Extended Conceptual Metaphor Theory

Extended Conceptual Metaphor Theory
The phenomenon known as metaphor is an extremely complex mental event –

we cannot capture its complexity if we tie ourselves to existing standard
views on metaphor. This book offers fresh insight into metaphor, updating an
established theory, conceptual metaphor theory (CMT), in the context of
current cognitive linguistic theory, and clarifying many of the issues that
researchers in the study of metaphor have raised against CMT. Starting with
an introduction to CMT, the subsequent chapters set out propositions for
extended conceptual metaphor theory, including a discussion on whether
literal language exists at all, whether conceptual metaphors are both concep-
tual and contextual, and whether they are both offline and online. Providing a
fresh take on a constantly developing field, this study will enrich the work of
researchers in areas ranging from metaphorical cognition to literary studies.
zoltán kövecses is Emeritus Professor at Eötvös Loránd University.

He is an associate editor of Cognitive Linguistics and Metaphor and Symbol.
Extended Conceptual
Metaphor Theory
Zoltán Kövecses
Eötvös Loránd University
University Printing House, Cambridge CB2 8BS, United Kingdom
One Liberty Plaza, 20th Floor, New York, NY 10006, USA
477 Williamstown Road, Port Melbourne, VIC 3207, Australia
314–321, 3rd Floor, Plot 3, Splendor Forum, Jasola District Centre,
New Delhi – 110025, India
79 Anson Road, #06–04/06, Singapore 079906
Cambridge University Press is part of the University of Cambridge.

It furthers the University’s mission by disseminating knowledge in the pursuit of
education, learning, and research at the highest international levels of excellence.
www.cambridge.org
Information on this title: www.cambridge.org/9781108490870
DOI: 10.1017/9781108859127
© Zoltán Kövecses 2020
This publication is in copyright. Subject to statutory exception
and to the provisions of relevant collective licensing agreements,
no reproduction of any part may take place without the written
permission of Cambridge University Press.
First published 2020
Printed in the United Kingdom by TJ International Ltd, Padstow Cornwall
A catalogue record for this publication is available from the British Library.
Library of Congress Cataloging-in-Publication Data
Names: Kövecses, Zoltán, author.
Title: Extended conceptual metaphor theory / Zoltán Kövecses.
Description: Cambridge ; New York, NY : Cambridge University Press, 2020. |
Includes bibliographical references and index.
Identifiers: LCCN 2019040754 (print) | LCCN 2019040755 (ebook) |
ISBN 9781108490870 (hardback) | ISBN 9781108796620 (paperback) |
ISBN 9781108859127 (epub)
Subjects: LCSH: Metaphor. | Cognitive grammar.
Classification: LCC P301.5.M48 K67 2020 (print) | LCC P301.5.M48 (ebook) |
DDC 401/.43018–dc23
LC record available at https://lccn.loc.gov/2019040754
LC ebook record available at https://lccn.loc.gov/2019040755
ISBN 978-1-108-49087-0 Hardback
Cambridge University Press has no responsibility for the persistence or accuracy
of URLs for external or third-party internet websites referred to in this publication
and does not guarantee that any content on such websites is, or will remain,
accurate or appropriate.
Very much belatedly. . ., to the memory of my
little sister, Annika.
Contents
List of Figures page viii

List of Tables ix
Preface xi
1 A Brief Outline of “Standard” Conceptual Metaphor

Theory and Some Outstanding Issues 1
2 The Abstract Understood Figuratively, the Concrete Understood
Literally, but the Concrete Understood Figuratively? 22
3 Direct or Indirect Emergence? 34
4 Domains, Schemas, Frames, or Spaces? 50
5 Conceptual or Contextual? 93
6 Offline or Online? 117
7 The Shape of the Extended View of CMT 150
8 By Way of Conclusion: Responses to the Five Questions 169
References 186
Index 194
Figures
3.1 Schematization of the prospective source domain page 40

3.2 Schematization of the prospective target 42
4.1 Schematicity hierarchy for four conceptual structures 52
4.2 Schematicity as inclusion 52
4.3 The schematicity hierarchy with three major distinctions
between the levels 55
4.4 Same-level mapping 58
4.5 The complex abstract systems are buildings metaphor 66
4.6 Activation and structuring between the levels 70
4.7 Activity is movement 76
4.8 Mark Tansey: Picasso and Braque (1992) 81
4.9 Complex abstract system is complex physical object 82
4.10 The movement image schema 84
4.11 The source-path-goal image schema 84
4.12 Levels of schematicity and methods of metaphor research 88
5.1 Summary of context types 101
5.2 The influence of context on schematicity hierarchies 105
5.3 Winner’s podium 109
5.4 Buda Castle 110
5.5 View of the Danube at Budapest 112
5.6 Dover Beach 114
6.1 Generic space and similarity in “This surgeon is a butcher” 135
6.2 The new surgeon as butcher blend in the “meaning
focus” view 141
7.1 Shared image schema 161
7.2 A process model for conceptual metaphors 167
8.1 Figurative construal and ontological content in concrete
and abstract concepts 170
8.2 Meaning construction for “The 2005 hurricane capsized
Domino’s life” 179
Tables
7.1 Summary of key ideas of the extended CMT view page 165
7.2 Four context types and their contextual factors 165
Preface
In a way, this is a very personal book. It is personal in the sense that in

everyone’s life there comes a time when you notice you are getting a little
impatient with how things are going and you want to say what’s really on your
mind concerning a subject matter that is important to you. I want to do exactly
that as regards metaphor. I do not want to be particularly concerned with what
other scholars will think or say when they read this book. I just want to lay my
(metaphorical) cards on the table. I am at an age when people around you die
much more frequently than before and when you realize it can happen any time
to you. I feel the time has come to make this move in the game of life.
This of course is not intended as a justification for writing another book on
metaphor. It can only explain some of the bold suggestions that go against
“standard” views on metaphor and the sometimes very straightforward ways of
writing and presenting ideas. What would be a much better justification,
though, is that over the past almost forty years I have analyzed thousands or
maybe even tens of thousands of metaphors in a variety of genres, modalities,
ways, and situations, and that these analyses have given me a certain feel for
how metaphors work and an idea of their enormous complexity. I realized, and
wish to show in a clear way here, that this complexity cannot be captured if
you tie yourself to existing standard views on metaphor, including conceptual
metaphor theory (CMT).
At the same time, as can be expected, my starting point is what I take to be a
“standard” version of CMT. I believe CMT is a theory that can provide
powerful and coherent explanations for a variety of aspects of metaphor. In
my judgment, no other theory is as comprehensive as CMT. It took almost
forty years for CMT to reach this stage. It’s been steadily developing thanks to
the many great scholars who played key roles in making it what it is today.
I see the present book as just another contribution to this line of development –
as an organic part of all the efforts that have been put into making it better.
I am making no claim that the ideas put forward in this book are all brand
new or that I am the only one who thought of them. But many of them are new
and, in addition, I feel I can reasonably claim that no one has put them together
as I have in this book. This overall new view can be characterized by five bold
but tentative propositions:
It may be that there is no literal language at all.
It may be that metonymies are, in a sense, “more primary” than primary
metaphors.
It may be that conceptual metaphors are hierarchically linked conceptual
structures on different levels of schematicity.
It may be that conceptual metaphors are not only conceptual but also
necessarily contextual.
It may be that conceptual metaphor is simultaneously an offline and online
phenomenon (i.e., it is not only offline).
I will be referring to the new view as “extended conceptual metaphor theory,”
or “extended CMT,” for short.
Structure of the Book

The propositions above are discussed in five chapters of the book under the
following corresponding chapter titles:
The abstract unaderstood figuratively, the concrete understood liter-
ally, but the concrete understood figuratively? (Chapter 2)
Direct or indirect emergence? (Chapter 3)
Domain, schema, frame, or space? (Chapter 4)
Conceptual or contextual? (Chapter 5)
Offline or online? (Chapter 6)
These five chapters are preceded by an introduction to CMT, or, more pre-
cisely, what I consider to be the “standard” view (the core) of CMT. The
discussion in the five chapters is followed by two integrative summary chap-
ters. The goal of Chapter 7 is to identify components of an emerging new
theory and to outline its general framework. The goal of Chapter 8 is to assess
the responses to the five questions above, together with a brief comparison of
the extended CMT with its sister theory, the dynamic systems view of meta-
phor, as proposed by Ray Gibbs.
Acknowledgments
No book is written in a vacuum. In addition to the vast body of published work
on conceptual metaphor theory, in particular, and on metaphor, in general, a
large number of colleagues, anonymous reviewers of this book, fellow
researchers, students, conference participants, and just ordinary people inter-
ested in my work helped me with their ideas and suggestions in many ways.
I am thankful to them all. Special thanks go to the following people (in
alphabetical order): Kathleen Ahrens, Valentina Bambini, Antonio Barcelona,
Réka Benczes, Bogusław Bierwiaczonek, Anna Borghi, Mario Brdar, Rita
Brdar-Szabó, Szilvia Csábi, Alice Deignan, John Douthwaite, Rachel Giora,
Andrew Goatly, Patrick Colm Hogan, Robert Hoffman, Robert Kardela, Sonja
Kleinke, Francisco Ruiz de Mendoza, Andreas Musolff, Uwe-Klaus Panther,
Frank Polzenhagen, Michele Prandi, Günter Radden, Elena Semino, Veronika
Szelid, Elzbieta Tabakowska, John Taylor, Linda Thornburg, and Cliff
Winters. I am grateful to all of them. Several of my students in Budapest
and Heidelberg have helped me shape the ideas in this book: Mohsen Bakhtiar,
Olga Boryslavska, Orsolya Farkas, Eszter Nucz, Orsolya Putz, and Erzsébet
Tóth-Czifra. Olga Boryslavska helped me with several of the figures in the
book. I am thankful to them for the many, sometimes passionate but always
enjoyable, discussions. I am grateful to Helen Barton at CUP for her encour-
agement and support.
Two people have played very special roles in the course of my career as a
metaphor researcher. Ray Gibbs has always been available to discuss various
issues related to the field and he and his work gave me a huge amount of
encouragement and inspiration. And last but definitely not least, without
George Lakoff I could not have and would not have done any of my work
on metaphor.
1 A Brief Outline of “Standard” Conceptual
Metaphor Theory and Some Outstanding Issues
Conceptual metaphor theory (CMT) started with George Lakoff and Mark
Johnson’s book, Metaphors We Live By (1980).1 The theory goes back a long
way and builds on centuries of scholarship that takes metaphor not simply as
an ornamental device in language but as a conceptual tool for structuring,
restructuring, and even creating reality. Notable philosophers in this history
include, for instance, Friedrich Nietzsche and, and more recently, Max Black.
A recent overview of theories of metaphor can be found in Gibbs (2008) and
that of CMT in particular in Kövecses (2002/2010).
Since the publication of Lakoff and Johnson’s (1980) work, a large amount
of research has been conducted that has confirmed, added to and also modified
their original ideas. Often, the sources of the new ideas were Lakoff and
Johnson themselves. Given this situation, it is obvious that what we know as
CMT today is not equivalent to the theory of metaphor proposed in Metaphors
We Live By. Many of the critics of CMT assume, incorrectly, that CMT equals
Metaphors We Live By. For this reason, I will not deal with this kind of
criticism in this introduction to CMT.
The standard definition of conceptual metaphors is this: A conceptual meta-
phor is understanding one domain of experience (that is typically abstract) in
terms of another (that is typically concrete) (see Lakoff and Johnson 1980).
This definition captures conceptual metaphors both as a process and a product.
The cognitive process of understanding a domain is the process aspect of
metaphor, while the resulting conceptual pattern is the product aspect. In this
survey of the theory, I will not distinguish between the two aspects.
In the following sections, I attempt to identify and briefly describe the main
features of CMT, as I see them. Other researchers might emphasize different
properties of the theory. At the same time, I tried to select those features
on which there is some agreement among practitioners of CMT. At the end of
the chapter and given the description of the properties, I list a number of
1
This chapter is a revised and enlarged version of a chapter I wrote entitled “Conceptual metaphor
theory” in Elena Semino and Zsófia Demjén, eds., 2017, The Routledge Handbook of Metaphor
and Language, 13–27. London: Routledge.
1
2 A Brief Outline of “Standard” Conceptual Metaphor Theory
outstanding issues in CMT – issues that we need to get clear about in order to
make CMT an even more powerful theory of metaphor.
1.1 The Pervasiveness of Metaphors

In their Metaphors We Live By, Lakoff and Johnson (1980) suggested that
metaphors are pervasive not only in certain genres striving to create some
artistic effect (such as literature) but also in the most neutral, i.e., most
nondeliberately used forms of language. CMT researchers, especially in the
early stages of work on conceptual metaphors, collected linguistic metaphors
from a variety of sources: TV and radio broadcasts, dictionaries, newspapers
and magazines, conversations, their own linguistic repertoires, and several
others. They found an abundance of metaphorical examples, such as
“defending an argument,” “exploding with anger,” “building a theory,” “fire
in someone’s eyes,” “foundering relationship,” “a cold personality,” “a step-
by-step process,” “digesting an idea,” “people passing away,” “wandering
aimlessly in life,” and literally thousands of others. Most, if not all, of such
linguistic metaphors are part of native speakers’ mental lexicon. They derive
from more basic senses of words and reflect a high degree of polysemy and
idiomaticity in the structure of the mental lexicon. The magnitude of such
cases of polysemy and idiomaticity in the lexicon was taken to be evidence of
the pervasiveness of metaphor. Based on such examples, they proposed what
came to be known as “conceptual metaphors.” However, it is still an issue
whether each and every metaphor we find in discourse belongs to a particular
conceptual metaphor. The answer to this question probably depends on the
level of schematicity at which we examine metaphors (see Chapter 4).
Other researchers, however, find the presence of metaphor in real discourse
less pervasive. As noted by Gibbs (2009), different methods produce different
results in frequency counts of metaphors.
1.2 A More Technical Definition of Conceptual Metaphors

The standard definition of conceptual metaphors we saw in Section 1.1 can be
reformulated somewhat more technically as follows: A conceptual metaphor is
a systematic set of correspondences between two domains of experience. This
is what “understanding one domain in terms of another” means. Another term
that is frequently used in the literature for “correspondence” is “mapping.”
This is because certain elements and the relations between them are said to be
mapped from one domain, the “source domain,” onto the other domain, the
“target.” Let us illustrate how the correspondences, or mappings, work with
the conceptual metaphor anger is fire. Before I provide the systematic
conceptual mappings that constitute this metaphor, let us see some linguistic
1.2 A More Technical Definition of Conceptual Metaphors 3
metaphors, first, as derived by the lexical method, then, as derived by corpus

studies, that make the conceptual metaphor manifest in English (the distinction
between lexical and corpus-based studies will be taken up in Chapter 4):
That kindled my ire.
Those were inflammatory remarks.
Smoke was coming out of his ears.
She was burning with anger.
He was spitting fire.
The incident set the people ablaze with anger.
We can find several of these linguistic metaphors in corpus-oriented work as

well, such as Charteris-Black’s most recent book (Charteris-Black 2017). Let
us take the following examples:
Therefore the wrath of the Lord was kindled against His people, So that He
abhorred His own inheritance. (Psalms 106:40) (Charteris-Black 2017: 80)
McInturff backed McCain’s decision to ignore the minister’s inflammatory
anti-America comments because it would have been seen as race-baiting
and sparked racial anger and protests. (Charteris-Black 2017: 44) (bold-
ing on “inflammatory” added, ZK)
Her ears were burning with rage. It was a commingling of pride, anger, pain
and frustration that determined what she was able to do in the next few
moments. (Charteris-Black 2017: 49)
I see him now – his eyes blazing forth with indignation and his rusty tousled
head of hair standing on end – leading forth on the miseries of the Gorbals
district and the East End of Glasgow: I was quite moved: I thought
everybody appreciated to the full the enthusiastic and fiery speech: The
whole passion of the man called out for justice to be handed out to the
working classes in the various parts of the city. (Nicholson, 18th June
1947) (Charteris-Black 2017: 164) (bolding on “blazing” added, ZK)
Given such examples, the following set of correspondences, or mappings, can

be proposed:
the cause of fire ➔ the cause of anger
causing the fire ➔ causing the anger
the thing on fire ➔ the angry person
the fire ➔ the anger
the intensity of fire ➔ the intensity of anger
With the help of these mappings, we can explain why the metaphorical
expressions listed earlier mean what they do: why, for instance, kindle and
inflammatory mean causing anger, and why burning, spitting fire, and being
ablaze with anger indicate a high intensity of anger, with probably fine
distinctions of intensity between them. As can be seen, the mappings are
provided at a general level in standard CMT. However, as, for example,
Charteris-Black (2017) points out, there can be subtle differences between, for
instance, the causes and intensities of target concepts that are indicated by
various specific lexical items. As an example, we can mention the words
kindle, spark, inflammatory, and so on, that refer to different kinds of causes.
The mappings as usually formulated are not sensitive to this level of
specificity. (But the way this can be remedied in CMT will be discussed in
Chapters 4–6).
The generic set of mappings is systematic in the sense that it captures a
coherent view of fire that is mapped onto anger: There is a thing that is not
burning. An event happens (cause of fire) that causes the fire to come into
existence. Now the thing is burning. The fire can burn at various degrees of
intensity.
Similarly for anger: There is a person who is not angry. An event happens
that causes the person to become angry. The person is now in the state of
anger. The intensity of the anger is variable.
The mappings bring into correspondence the elements and the relations
between the elements in the fire domain (source) with elements and the
relations between the elements in the anger domain (target). Indeed, it seems
reasonable to suggest that, in a sense, the mappings from the fire domain
actually bring about or create a particular conception of anger relative to the
view of fire we have just seen. This is what it means that a particular source
domain is used to conceptualize a particular target domain. (I will come back
to this issue later.)
In many cases, however, the two-domain account does not work and must
be supplemented by a model of explanation that relies on four (or more)
domains, or spaces (see Chapter 6 on conceptual integration and metaphor).
1.3 Mapping Further Knowledge

Given the metaphorically used set of elements in a domain, we can derive
further knowledge about these elements, and can also map this additional
knowledge onto the target. This additional kind of source-domain knowledge
is often called “metaphorical inference,” or “metaphorical entailment.” For
example, to stay with the aforementioned metaphor, in somewhat formal and
old-fashioned English we can find sentences like “He took revenge and that
quenched his anger.” Quenching anger can be regarded as a metaphorical
inference, given the anger is fire metaphor. If anger is metaphorically
viewed as fire, then we can make use of our further knowledge of anger-as-
fire; namely, that the fire can be quenched. CMT provides an elegant explan-
ation of such cases of extending conceptual metaphors.
At this point, an important question may arise: Can everything be mapped
from one domain to another? Obviously, not. Given a particular conceptual
1.4 Mappings Go from Concrete to Abstract Domains 5
metaphor, there are many things that cannot be mapped, or carried over, from
the source to the target. For example, given that theories are buildings,
the number of rooms or whether the building has a cellar or an attic are not
mapped. Several explanations have been offered to delimit the amount of
knowledge that can be transferred from the source. One of them is the
“invariance hypothesis” developed by Lakoff (1990). It suggests that every-
thing from the source can be mapped onto the target that does not conflict with
the image-schematic structure of the target. Another is proposed by Grady
(1997a,b), who claims, in essence, that those parts of the source domain can be
mapped that are based on “primary metaphors” (on “primary metaphors,” see
Grady (1997a,b)). Finally, Kövecses (2000a, 2002/2010) proposed that the
source maps conceptual materials that belong to its main meaning focus or
foci. It should be noted that the three suggestions differ with respect to which
part of a conceptual metaphor they rely on in their predictions concerning what
is mapped. The first one relies primarily on the target, the second on the
connection between source and target, and the third on properties of the source.
None of these are entirely satisfactory.
1.4 Mappings Go from Concrete to Abstract Domains

As we just saw, CMT makes a distinction between a “source domain” and a
“target domain.” The source domain is a concrete domain, while the target is
an abstract one. In the example conceptual metaphor life is a journey, the
domain of journey is much more concrete than the target domain of life (that is
much more abstract); hence, journey is the source (domain). In general,
CMT proposes that more physical domains typically serve as source domains
for more abstract targets, as in the life is a journey metaphor.
This observation is based on the examination of hundreds of conceptual
metaphors that have been discovered and analyzed in the literature so far (such
as life is a journey, anger is fire, and theories are buildings).
The assumption that most conceptual metaphors involve more physical
domains as sources and more abstract domains as targets makes a lot of
intuitive sense. For example, the notion of life is hard to pin down because
of its complexity, that of anger is an internal feeling that remains largely
hidden from us, that of theory is a sophisticated mental construct, and so on
for other cases. In all of them, a less tangible and thus less easily accessible
target concept is conceptualized as and from the perspective of a more tangible
and thus a more easily accessible source concept.
In our effort to understand the world, it makes a lot more sense to move
conceptually in this particular direction: that is, to conceptualize the cogni-
tively less easily accessible domains in terms of the more easily accessible
ones. Notice how odd and unintuitive it would be to attempt to conceptualize
journeys metaphorically as life, fire as anger, or buildings as theories. We

would not find this way of understanding journey, fire, or building helpful or
revealing, simply because we know a lot more about them than about such
concepts as life, anger, or theory. This is not to say that the reverse direction of
conceptualization never occurs. It may occur, but when it does, there is always
some special poetic, stylistic, aesthetic, and so on, purpose or effect involved.
The default direction of metaphorical conceptualization from more tangible to
less tangible applies to the everyday and unmarked cases. Furthermore, even
concrete source domain concepts (such as fire) can be metaphorically under-
stood. This latter issue will be taken up in the next chapter.
1.5 Metaphors in Thought

According to CMT, metaphor resides not only in language but also in thought.
We use metaphors not only to speak about certain aspects of the world but also
to think about them. As we saw earlier, CMT makes a distinction between
linguistic metaphors, i.e., linguistic expressions used metaphorically, and
conceptual metaphors, i.e., certain conceptual patterns we rely on in our daily
living, to think about aspects of the world. For example, metaphors such as
life is a journey can actually govern the way we think about life: we can
set goals we want to reach, we do our best to reach those goals, we can make
careful plans for the journey, we can prepare ourselves for facing obstacles
along the way, we can draw up alternative plans in the form of choosing from
multiple paths, we can prefer certain paths to others, and so on. When we
entertain such and similar ideas, we actually think about life in terms of the
life is a journey conceptual metaphor. And, consequently, we can use the
language of journeys to also talk about life.
The idea that we think about a domain in terms of another can actually mean
several different things. In one sense, as described previously, people may be
guided by a particular conceptual metaphor in how they conceive of a domain,
such as life. In another, given a conceptual metaphor, they may utilize some of
the implications of a particular domain they rely on (such as journey) in a
conceptual metaphor and apply those implications to the other domain (such as
life) in their reasoning about it (see below for an example). Finally, it can also
mean that in the course of the online process of producing and understanding a
linguistic metaphor, the metaphor activates both the source and the target
concept. (This issue is discussed in the final two chapters of the book.)
1.6 Metaphor and the Construction of Reality

A major consequence of the idea that metaphors are conceptual in nature,
i.e., that we conceive of certain things in metaphorical ways, is that, since
1.7 Multimodality of Metaphors 7
our conceptual system governs how we act in the world, we often act
metaphorically.
When we conceptualize an intangible or less tangible domain metaphoric-
ally as, and from the perspective of, a more tangible domain, we create a
certain metaphorical reality. We imagine life one way when we think of it as a
journey (see the aforementioned discussion), and in another way when we
think of it as a theatre play, as reflected in Shakespeare’s famous lines “All the
world is a stage / and all men and women are merely players.” The two source
domains result in very different views on life, and in this sense they create very
different realities.
Whenever a new source domain is applied to a particular target, we see the
target domain differently than we saw it before. The limiting case of this
situation is the one when a particular target domain does not exist at all, but
by the application of one (or several) source domain(s), it actually gets created.
Very often, the etymologies of words for abstract concepts reflect this early
conceptualization. For example, comprehension (“understanding”) is
clearly an abstract concept. Given the understanding is grasping con-
ceptual metaphor (as in “I did not grasp what he said,” “He is slow on the
uptake”), it makes sense that the English word comprehend derives from the
word that means “grasp” in Latin.
This kind of “reality construction” is very common in advertising, where,
often, interesting or amusing cases of metaphorical reality get created. When
advertisements for, say, deodorants promise “24-hour protection,” they make
us see a deodorant as our helper or ally in a fight or war against an enemy. The
enemy is no other than our own body odor. So if we did not think of our body
odor as our enemy before, i.e., as something we have to be protected against,
the advertisements can easily make us view it as such. In this manner, the
metaphors used in advertisements and elsewhere can create new realities for
us. Such realities are of course metaphorically defined. But this does not make
them unimportant for the way we live. If we think of our body odor as
something we need to be protected against and as a result go and buy a
deodorant to overcome the enemy, we are clearly thinking and acting
according to a metaphorically defined reality. This is a further example of
how the implications of a source domain for a particular target can be utilized
(in a process I called metaphorical inference or entailment in the previous
section).
1.7 Multimodality of Metaphors

Finally, if metaphor is part of the conceptual system, it follows that conceptual
metaphors will also occur in any mode of expression of that system. Research
indicates that the conceptual metaphors identified in language also occur in
gestures, visual representations (such as cartoons), visual arts (such as

painting), and others. This does not mean that the metaphors found in these
modes of expression are exactly the same as found in everyday language and
thought, but that a large number of them are (see, e.g., work by Cienki and
Müller 2008; El Refaie 2019; Forceville 2008, 2016; Lakoff 1993).
1.8 The Grounding of Conceptual Metaphors

Why is a particular source domain paired with a particular target domain? The
most traditional answer to this question is to say that there is a similarity, or
resemblance, between two things or events. Several different types of similar-
ity are recognized in the literature: objectively real similarity (as in the roses on
one’s cheeks), perceived similarity, and similarity in generic-level structure.
An example for perceived similarity would be a case where certain actions in
life and their consequences are seen as gambles with a win or lose outcome in a
gambling game; cf. life is a gambling game. We can take as an example
for the last type of similarity the conceptual metaphor human life cycle is
the life cycle of a plant. The two domains share generic-level structure
that can be given as follows: In both domains, there is an entity that comes into
existence, it begins to grow, reaches a point in its development when it is
strongest, then it begins to decline, and finally it goes out of existence. Based
on this shared structure, the plant domain can function as a source domain for
the human domain. In other words, the similarity explains the pairing of this
particular source with this particular target; that is, the metaphor is grounded in
similarity – though of a very abstract kind.
In many other cases, however, this explanation does not work: The source
cannot be viewed as similar in any way to the target. CMT offers another
explanation or justification for the emergence of these metaphors as well. Let
us take the conceptual metaphor in one of the metaphor systems we examined
in the previous section: intensity is heat. This metaphor is a generic-level
version of a number of conceptual metaphors like anger is fire, enthusi-
asm is fire, conflict is fire, and so on. The specific concepts share an
intensity dimension that is metaphorically conceptualized as heat. The concept
of heat bears no resemblance to that of intensity whatsoever. Heat is a
physical property of things that we experience with our bodies, while intensity
is a highly abstract subjective notion (on a par with purpose, difficulty, or as a
matter of fact, similarity). What, then, allows the use of heat as a source
domain for intensity? CMT suggests that there is a correlation in experi-
ence between intensity and heat. Often, when we engage in activities at a high
intensity (be it physical or emotional), our body develops body heat. In this
sense, intensity is correlated with heat, and this provides the motivation for the
use of heat as a source domain for intensity as a target. The generic-level
1.11 Metaphor and Grammar 9
conceptual metaphor intensity is heat can then be regarded as grounded

in a correlation in experience between a sensory-motor experience and an
abstract subjective one.
1.9 Primary and Compound Metaphors

Conceptual metaphors of this kind are called “primary metaphors” by Lakoff
and Johnson (see, e.g., 1999), who borrowed the term and idea from Joe Grady
(1997a,b) and developed it further. Grady proposed a number of such meta-
phors in his dissertation (1997a), including similarity is closeness,
persistence is being erect, and reanalyzed several of the conceptual
metaphors in Lakoff and Johnson’s early work (1980) along the same lines
(e.g., more is up, purposes are destinations). He suggested further-
more that several primary metaphors can be put together to form “compound
metaphors.” For example, the purposeful life is a journey metaphor is
based on the primary metaphors purposes are destinations, difficul-
ties are impediments, and others.
1.10 Image Schemas and Metaphors

Many conceptual metaphors (both the similarity-based ones and the primary
metaphors) are based on “image schemas.” These are abstract, preconceptual
structures that emerge from our recurrent experiences of the world (Johnson
1987; Lakoff 1987). Such skeletal preconceptual structures include con-
tainer, source-path-goal, force, verticality, and several others.
For example, the states are containers primary metaphor derives from
the container image schema, the life is a journey metaphor from the
source-path-goal schema, the emotions are forces metaphor from
the force schema, and so on.
1.11 Metaphor and Grammar

The main aim of CMT is to analyze and describe the conceptual nature of
metaphor: how conceptual metaphors structure thought, how they enable
inferences, how they can give us new perspectives on reality, how they can
construct new ideas and concepts, how they are grounded in experience, and so
on. However, metaphor scholars have increasingly realized in recent years that
there is a close connection between language structure (i.e., grammar) and
metaphorical conceptual structure. We are beginning to see that metaphor and
grammar (Langacker 1987; Goldberg 1995) can cross-fertilize each other.
Let us just look at a single problem that can demonstrate this. Sullivan
(2016) poses the question: Why is it that, given the sentences “The lawyer
devised an argument,” “The lawyer devised a house,” “The carpenter devised

an argument,” and “The lawyer constructed/built an argument,” only the last
one (with the verb construct or build) is metaphorical? The first sentence
describes literally what lawyers do. The other three, however, should all be
instances of the theories are buildings metaphor, since they contain
words that have to do with the source domain of buildings: house, carpen-
ter, and construct/build. After all, just like the building of the house corres-
ponds to the creation of the argument and is thus metaphorical, the house
should correspond to the argument and the carpenter should correspond to the
person making the argument, but they don’t and thus they are not metaphor-
ical. The answer to the question is that, given the transitive construction with
an agent, verb, and patient, only the verb (construct or build) is metaphorical in
the sentences because it is a dependent element of the construction, while the
other two (the two nouns) are autonomous within a cognitive grammar frame-
work. As Sullivan (2016: 145) notes, “Elements are dependent when their
meanings are incomplete without one or more autonomous elements.” More
generally, dependent elements in a construction evoke source domains (thus,
are metaphors) and the autonomous elements in constructions evoke target
domains (thus, are literal).
The study of the interaction between metaphor and (cognitive) grammar is a
promising new avenue in CMT. For example, results such as the above make
sense of corpus findings according to which “verbs are more frequently used
metaphorically (that is, to evoke metaphoric source domains) than nouns”
(Sullivan 2016: 143), as observed, for instance, by Cameron (2003) and
Deignan (2005) in their corpus-based studies.
1.12 The Neural Theory of Metaphor

The research on primary metaphors has intensified the study of metaphors in
the brain. Lakoff (1993) suggested a “neural theory of metaphor.” In it,
individual neurons in the brain form neuronal groups, called “nodes.” There
can be different types of neural circuits between the nodes. In the “mapping
circuit” that characterizes metaphor, there are two groups of nodes correspond-
ing to source and target domain. The circuitry between the two groups of nodes
will correspond to the mappings, or correspondences. In primary metaphors,
one group of nodes represents a sensorimotor experience in the brain, while the
other represents an abstract, subjective experience. If the neural theory of
metaphor is correct, it leads to an important conclusion concerning our func-
tioning in the world: we do not only understand (or conceptualize or think)
about target domains in terms of source domains, but we experience target
domains as source domains. This is the foundational idea of accounts of
metaphor processing based on perceptual simulations.
1.15 Metaphor Networks 11
1.13 Universality and Variation

Since the human body and the brain are predominantly universal, the meta-
phorical structures that are based on them will also be predominantly universal.
This explains why many conceptual metaphors, such as knowing is seeing,
can be found in a large number of genetically unrelated languages. This does
not mean, however, that all conceptual metaphors that are based on primary
metaphors will be the same from language/culture to language/culture. It was
recognized early on that the particular culture in which a metaphor develops is
just as significant in shaping the form of the conceptual metaphors in different
languages/cultures as the universal bodily experiences themselves (see, e.g.,
Kövecses 1995b; Musolff 2004; Taylor and MacLaury 1995; Taylor and
Mbense 1998; Yu 1998, 2002;). Furthermore, several researchers pointed out
that variation in metaphor can also be found within the same language/culture
(for a survey of this research, see, e.g., Kövecses 2005).
1.14 Context and Metaphor

As the latest development in this trend, scholars have recognized that it is not
only culture that functions as an important kind of context in shaping the
metaphors that emerge. More and more researchers in this area take into account
the tight connection between metaphorical aspects of our cognitive activities
and the varied set of contextual factors that influence the emergence of meta-
phors (see, e.g., Cameron 2003; Gibbs and Cameron 2008; Goatly 2007;
Kövecses 2010a,b; Semino 2008;). The overall result is a much richer account
of metaphor. First, it has become possible to account for metaphors that may be
completely everyday but at the same time do not fit any pre-established concep-
tual metaphors (see, e.g., the work by Musolff 2004; Semino 2008; and others).
Second, by taking into account the role of context, we are now in a much better
position to see a fuller picture of metaphorical creativity than before. Indeed, it
can be suggested that contextual factors can actually create novel metaphors that
can be referred to as “context-induced” ones (Kövecses 2010b, 2015a). Third,
these context-induced metaphors are not limited to the kinds of basic correl-
ations in experience that form the bases of primary metaphors. Thus, we seem to
have a cline of metaphors, ranging from universal primary metaphors to non-
universal context-induced ones. In other words, metaphors can derive from the
body, cultural specificities, and also the more general context.
1.15 Metaphor Networks

The notion of “network” has been recently applied to conceptual metaphors
and metaphor-related phenomena. Most famously, it was used by Fauconnier
and Turner (e.g., Fauconnier and Turner 2002) in their discussion of concep-
tual integration, where they assume a “network model” that consists of
four spaces (see Chapter 6 for details). Richard Trim (2007, 2011) uses the
term “metaphor network” for the development of an individual metaphor,
such as the lion as a symbol of courage in British heraldry, as well as for a
network of metaphors that grows historically from a core concept, such as
mother or father – giving rise to such metaphors as mother tongue or
mother country in English. He shows how such networks emerge in a number
of European languages and cultures over the ages. Needless to say, both of
these are extremely interesting and important applications of the idea of
“network,” but I have an application of the term in mind that is different
from both.
As we have seen above, the source domains of conceptual metaphors
constitute coherent organizations of experience, and the mappings from the
source onto the target domains create equally systematically organized target
domains. But the question is whether such systematic source to target map-
pings are isolated from each other. I suggest that they are likely to belong to
larger, hierarchically organized systems of metaphors.
The principles for the organization of such metaphor systems can be of
several distinct kinds. In one (a), the metaphors are organized in a straightfor-
ward hierarchy such that both the source and the target are specific cases of
higher generic-level concepts. In another (b), different aspects of a given
generic-level concept can be differentially conceptualized by means of con-
ceptual metaphors. In still other cases (c), a single aspect of several different
abstract concepts may organize a large number of subordinated specific-level
conceptual metaphors into a hierarchy. In a fourth (d), the conceptual meta-
phors form a system because the target domains are part of an independently
existing hierarchy of concepts. In a fifth (e), what connects the conceptual
metaphors and makes them form a system is the fact that a particular specific
level target concept is a special case of a number of different higher-level
concepts that have their own characteristic conceptual metaphors. There are
probably additional ways in which metaphor systems (or networks) are
formed, but for the present purposes it is sufficient to take these five possibil-
ities into account and briefly describe them.
Straightforward Hierarchies
In this case, both the source and the target are specific-level concepts of
generic-level conceptual metaphors. This is the simplest and most straightfor-
ward type of hierarchy, and it involves a large number of cases. Let us take the
well-known anger is a hot fluid metaphor. This is an instance of the
generic-level metaphor emotions are forces. Actually, the hot fluid
source can be further specified, yielding, for example, the concept of stew as
a potential source domain. We can represent this as follows:
emotions are forces
anger is a hot fluid in a container (He was boiling
with anger.)
anger is a stew (He was stewing.)
We can find the same situation for love:
emotions are forces
love is a natural force (I was overcome with love.)
love is the wind (It was a whirlwind romance.)
As we will see throughout this book, despite, or rather, because of, its simplicity,
this kind of network is crucially important for the extended theory of metaphor
I am going to present in this work. I will come back to it in several later chapters.
Different Aspects of a Single Generic Concept

What is known as the Event Structure metaphor system presents a more
complicated situation (see Lakoff 1993). Events in general (i.e., the generic-
level concept of event) can be actions and occurrences, and they both involve
states, causes and changes. Actions also include long-term activities, where
progress is an issue. Actions are characterized by purposes, potential difficulty
in execution, and manner of performance. These various aspects of events
(event 1) are conceptualized in different ways:
Events: events are movements
Occurrences: occurrences are movements (What’s going on here?)
Actions: actions are self-propelled movements (What’s going to
be the next step?)
Cause: causes are forces (You’re driving me nuts.)
State: states are locations / bounded regions (She’s in love.)
Change: change is motion (from one location to another)
(I almost went crazy.)
Actions: actions are self-propelled movements
Purpose: purposes are destinations (I want to reach my goals.)
Difficulty: difficulties are impediments (to motion) (Let’s get
around this problem.)
Manner: manner is path (of motion) (We’ll do it in another way.)
Activity: long-term purposeful activities are journeys (We have
a long way to go with this project.)
Progress: expected progress is a travel schedule (I am way behind
schedule.)
A further conceptual metaphor for progress is exemplified in the Activity
component above: progress is distance covered (to a destination)
(as in “have a long way to go”).
As we can see, the highest-level metaphor here is related to the overarching

category of events: events are movements. Events come in several forms,
and they are characterized by several different aspects. The various forms and
aspects of events are in turn metaphorically viewed in terms of the source
domains of movement, location, and force. These can of course be further
elaborated at still more specific levels of concepts.
A Single Aspect of Several Different Specific-Level Concepts

Several conceptual metaphors may belong together by virtue of the fact that
they share a particular aspect that is conceptualized metaphorically by means
of the same source domain. The target domains to which a single source
domain is applied is the “scope of a source domain” (Kövecses 2000a,
2010a,b). Thus, the scope of a source can be narrow or wide. Consider the
following conceptual metaphors:
anger is fire (He was smouldering with anger.)
love is fire (The fire was gone from their relationship.)
desire is fire (It was his burning ambition to become a lawyer.)
imagination is fire (The scene set fire to his imagination.)
enthusiasm is fire (He lost the fire.)
conflict is fire (The fire of war burnt down Europe several times in the
course of its history.)
energy is fire (She’s burning the candle at both ends.)
All of these target domains share the aspect of (degrees of ) intensity through
the application of a single source (heat of fire). We can suggest that the
fire source domain has the “main meaning focus” of intensity (Kövecses
2000a, 2002/2010). Thus, one way of metaphorically understanding intensity
is in terms of the heat of fire. This yields the generic metaphor intensity is
heat. Consequently, the specific metaphors above are instances of this
generic-level metaphor. This is a further way in which conceptual metaphors
may form a hierarchical system. As a matter of fact, primary metaphors (see
above) can be seen as forming such systems in a natural way, since their target
domains represent shared aspects (like intensity) of several different concepts.
Several Different Aspects of a Single Specific-Level Concept

A specific-level abstract concept may inherit conceptual metaphors from
several different generic-level metaphor systems by virtue of the fact that its
prototypical cognitive-cultural model consists of elements that belong to the
different metaphor systems. We can exemplify this with the specific-level
abstract concept of friendship (Kövecses 1995a). The model of friendship
conceptually partakes of a number of different metaphor systems. Since

according to the cognitive-cultural model of friendship,
it is a state that two people attribute to each other,
it involves communication between the friends,
it implies mutual interaction with each other,
it consists of the friends and their interactions as a complex system,
it includes participants that feel certain emotions towards each other,
and some others,
the conceptual metaphors that characterize friendship include the following:
State metaphor system:
states are objects
attributed states are possessed objects (Lakoff 1993)
Communication metaphor system:
the mind is a container
linguistic expressions are containers
meanings are objects
communication is sending (Reddy 1979)
Interaction metaphor system:
interactions are economic exchanges (Kövecses 1995a)
Complex system metaphor system:
abstract complex systems are complex physical systems
(Kövecses 1995a, 2010a)
Emotion metaphor system:
emotion is distance
emotion is temperature (Kövecses 1990, 2000b)
The conceptual metaphors for friendship emerge from these various metaphor
systems. Specifically, we find metaphors such as the following in the descrip-
tions of friendship:
State metaphor system:
friendship is a possessed object (My friendship with her did not
last long.)
Communication metaphor system:
sharing (communicating) experiences is sharing objects:
(We share intimate things with each other.) (Kövecses 1995a, 2000b)
The metaphor arises because communication between friends often involves
sharing ideas and feelings.
Interaction metaphor system:
interactions in friendship are economic exchanges (There is a
lot of give and take in our friendship.) (Kövecses 1995a)
The interactions are conceptualized as “economic” exchanges because people

often mention a fifty-fifty basis in their friendship interactions, which indicates
not just a physical exchange of objects.
Complex system metaphor system:
friendship is a complex physical system (building, machine,
plant) (We have built a strong friendship over the years.) (Kövecses
1995a, 2002/2010)
Emotion metaphor system:
an emotional relationship is a distance (They have a close
friendship.)
emotion is temperature (They have a warm friendship.) (Kövecses
1990, 2000b)
The last two conceptual metaphors have to do with the notion of intimacy that
characterize several emotions, yielding the metaphors intimacy is close-
ness and intimacy is warmth (both of which are primary metaphors).
More generally, since aspects of friendship constitute a part of these meta-
phor systems, the hybrid concept of friendship will share with them the
specific metaphors.
The Target Concepts Form a Hierarchical System of Concepts

The best example for this kind of metaphor system is what is called the Great
Chain of Being (Lakoff and Turner 1989). This is a hierarchical system of
concepts corresponding to objects and entities in the world, such as humans,
animals, physical things, and so on. The extended version of this hierarchy
consists of the following (Lakoff and Turner 1989; Kövecses 2002/2010):
God
Complex systems (universe, society, mind, theories, company, friend-
ship, etc.)
Humans
Animals
Plants
Complex physical objects
Inanimate objects
The hierarchy becomes a metaphor system when things on a particular level
are conceptualized as things on another level. Notice that this can happen in
both directions. Lower-level concepts can function as source domains for
higher-level ones as target (e.g., people are animals) and higher-level
ones can function as source domains for lower-level ones as target (e.g.,
animals are people). Furthermore, the human, animal and plant
1.16 Critique of CMT 17
categories are often graded internally – a conceptualization that can lead to

racist language (e.g., an “inferior race”).
In summary, conceptual metaphors are not isolated conceptual patterns in
the mind but seem to cluster together to form a variety of interlocking
hierarchical relationships with each other.
1.16 Critique of CMT

In the world of academia, CMT is in a curious position: despite its many
undeniable achievements, its obvious usefulness in and popularity across
several disciplines, each and every aspect of it has come under criticism in
the past thirty years (see, e.g., Fusaroli and Morgagni 2009; Gibbs 2017a).
Indeed, several scholars have expressed their skepticism regarding the very
existence of conceptual metaphors (e.g., Cameron and Maslen, eds. 2010).
A further curious aspect of the situation is that a considerable body of the
criticism is based on Lakoff and Johnson’s (1980) work exclusively, which
represents only the initial stage of CMT, ignoring much of the later work in
CMT. Since this chapter has described, or at least briefly mentioned, some of
that work subsequent to Metaphors We Live By, I will not take up the criticisms
that relate to these features of CMT.
A charge that is sometimes levelled at CMT is that it works with the concept
of domain (as in the idea that conceptual metaphors involve two domains) and
that is itself not a well-defined concept and that it probably cannot be defined
precisely at all. Indeed, CMT works with a number of different conceptions
that rely on the idea of a coherent organization of human experience, such as
domain, frame, image schema, and mental space. This situation presents a
challenge to conceptual metaphor theory.
Another criticism maintains that CMT is based on circular reasoning. Here
the claim is that scholars in CMT use linguistic metaphors to identify concep-
tual metaphors, on the one hand, and that at the same time they suggest that the
linguistic metaphors exist because of the already present conceptual ones, on
the other. One cannot justify the existence on conceptual metaphors on
linguistic metaphors and at the same time explain the presence of linguistic
metaphors on the basis of conceptual metaphors. However, this criticism
ceased to be valid after several experiments that did not involve language or
linguistic metaphors (beginning with Gibbs’ work in the early 1990s) unam-
biguously confirmed the existence of conceptual metaphors. If conceptual
metaphors have been proven to have psychological reality by psycholinguistic
experiments, linguists should not deny their existence; they should work to see
how they appear and function in language (and other modalities). (For sum-
maries of these experiments, see Gibbs 1994, 2006; Gibbs and Colston 2012.)
But the most commonly and strongly expressed criticism concerns meth-
odological issues; namely, how to identify metaphors in discourse, how the
study of metaphor should be based on real data (rather than just lexical or
intuitive data), and others (see, e.g., Deignan 2005; Pragglejaz Group 2007).
As I indicated above, we should now take these developments as an integral
part of CMT. However, the issue of the need to use real data for metaphor
analysis reveals an apparently real weakness of CMT: it is that CMT
researchers do not pay sufficient attention to the discourse and social-
pragmatic functions of metaphor in real discourse. This sounds like a valid
point. However, I do not think that CMT should be thought of as a view of
metaphor whose only job is to collect metaphorical expressions, set up
conceptual metaphors based on the expressions, lay out the mappings that
constitute those conceptual metaphors, and see how the particular conceptual
metaphors form larger systematic groups. A large further part of the mission
of CMT is to describe the particular syntactic, discursive, social, pragmatic,
rhetorical, aesthetic, etc., behavior and function of the metaphors in real data.
And this is precisely something that is currently conducted by a great number
of researchers (e.g., Low et al. 2010). But, to my mind, these researchers
are not competing with more “traditional” CMT scholars; instead, they are
working out an aspect of CMT that was “neglected” by CMT scholars. The
addition is necessary and more than welcome. This kind of work is just as
much part of CMT as other aspects of the theory. In other words, I find that
the “neglect” was not really neglect. The lack of sufficient attention to the
syntactic, pragmatic, etc., features of metaphors resulted from CMT scholars’
effort to add a cognitive dimension to metaphor that was mostly lacking in
previous work. This was, and still is, a major part of CMT scholars’ mission,
in collaboration with other metaphor researchers. Without pursuing that
mission, we would not have been able to link the study of metaphor with
the study of cognition in general.
1.17 The Present State of CMT

In my view, CMT is a complex and coherent theory of metaphor. As even the
sketchy picture above reveals, CMT is a theory of metaphor that is capable of
explaining a variety of issues concerning metaphor. In particular, it can
explain:
why we use language from one domain of experience systematically to talk
about another domain of experience;
why the polysemy of words in the lexicon follows the patterns it does;
why the senses of words are extended in the concrete-to-abstract direction;
why children acquire metaphors in the sequence they do;
why the meanings of words emerge historically in the sequence they do;
1.18 Extending CMT 19
why many conceptual metaphors are near-universal or potentially universal;

why many other conceptual metaphors are variable cross-culturally and
intraculturally;
why many conceptual metaphors are shared in different modes of expression
(such as verbal and visual);
why many metaphor-based folk and expert theories of a particular subject
matter are often based on the same conceptual metaphors;
why so many conceptual metaphors are shared between everyday language
and literature (and other forms of non-everyday uses of language);
why and how novel metaphors can, and do, constantly emerge; etc.
No other theory of metaphor is capable of explaining all of these issues (for a
comprehensive picture, see, Gibbs, ed. 2008). This does not mean, however,
that CMT has achieved a “state of perfection,” and that is has no room to
develop further.
1.18 Extending CMT

Even the sketchy survey above reveals a number of problems and weaknesses
inherent in CMT. These include at least the following:
(a) It is reasonable to problematize even a basic assumption of CMT: the
idea that there is such a thing as literal meaning. Is it indeed the case that
we can and do base our understanding of figurative meaning on literal
meaning? Even more radically: What is literal meaning? And, in the final
analysis, how would the answers to these questions impact CMT as we
know it today?
(b) If primary metaphors assume a single conceptual unit (a scene) with
elements that are correlated (such as anger and body heat), wouldn’t it be
justifiable to think of the emergence of primary metaphors through a
metonymic stage (in which, for example, anger and body heat are
correlated within a tightly organized structure), and not as emerging
directly as metaphors from two experiential “domains”?
(c) Indeed, what is the appropriate conceptual level at which we can, or
should, identify conceptual metaphors? Is it that of domain, frame, scene,
schema, space, or something else? Most of these have been mentioned in
the preceding discussion, but there is no consensus, even within the same
author, which one it is.
(d) And if we successfully clarify the issue of the appropriate conceptual
level for conceptual metaphors, what would be the implications of this?
Would there be any significant repercussions for CMT at all? For
example, would it change the way we think about the methodology of
studying metaphor?
(e) Several metaphor networks have been noted above. Are all of them
equally significant? Does any one of them stand out as especially import-
ant for the overall description of the metaphorical conceptual system?
(f ) How can the recently proposed view of deliberate metaphors be incorpor-
ated into a CMT framework? Or does it present a challenge to CMT?
(g) How can we account for the socio-pragmatic function of metaphorical
linguistic expressions in naturally occurring discourse? What would a
theory of conceptual metaphors look like that is capable of explaining the
conceptual structure of metaphors, as well as the discourse functions of
metaphorical expressions realizing them within a unified framework?
(h) Indeed, in connection with the previous question, should we think of
metaphor as a cognitive process that is happening online or as a product
that is somehow a part of a more or less stable conceptual system in long-
term memory?
(i) Though it is true that theories of metaphor comprehension make exten-
sive use of the notion of context, theories of metaphor production within
CMT have almost completely ignored the notion. Is it possible to propose
a theory of conceptual metaphors within the cognitive linguistic para-
digm that would integrate the context of metaphor production into CMT?
(j) The major idea of CMT is embodiment on which primary metaphors are
based. While clearly crucial, one can legitimately raise questions about
the exclusivity of embodiment in CMT. Isn’t it conceivable that there are
additional factors that play an important role in the creation of meta-
phors – both conceptual and linguistic?
(k) Concerning metaphorical creativity, CMT offers sophisticated accounts
of the novel metaphors in both every day and literary discourse. How-
ever, these accounts of metaphorical creativity have their limitations:
they can only explain creativity that is the result of putting together
potentially universal primary metaphors in unique ways or as a result
of a small number of conceptual devices (extending, elaboration, ques-
tioning, combining). However, in many cases we find novel metaphors
(both conceptual and linguistic) that require us to take into consideration
a variety of contextual factors, and not just universal body-based meta-
phors or universal cognitive processes.
(l) If we think with the help of conceptual metaphors, as we like to claim, it
is reasonable to ask why we use mixed metaphors in the actual produc-
tion of discourse. Wouldn’t it be easier and more efficient to use the same
conceptual metaphor throughout a given discourse? As a matter of fact,
metaphorically homogeneous discourse seems to be much rarer than
metaphorically heterogeneous discourse (i.e., mixing metaphors in dis-
course). Why is this the case? Can we provide a coherent CMT account
of mixed metaphor in discourse?
1.18 Extending CMT 21
(m) Cognitive linguists (including the present author) like to think of meta-
phors as “conceptual metaphors,” ignoring the diversity of phenomena
we call metaphor. What other kinds of metaphor are there, how can they
be characterized, and what is their relationship to conceptual metaphors?
Undoubtedly, there are many more questions that could be raised concerning
CMT. However, I find the questions in a. through m., among possibly others,
inevitable to raise in order to produce an extended, more comprehensive, and
hopefully improved, version of CMT. I address these issues in the following
chapters of the book.
2 The Abstract Understood Figuratively, the
Concrete Understood Literally, but the
Concrete Understood Figuratively?
Most metaphor researchers accept, either explicitly or implicitly, a distinction

between literal and figurative meaning in some form and base the distinction
between concrete and abstract concepts on it. But the distinction between the
figurative and the literal has also been problematized by a number of other
scholars (see the discussion in Katz et al. 1998). The issue is relevant to
conceptual metaphor theory in several ways and for several reasons – most
importantly because CMT relies heavily on the abstract-concrete distinction. In
this chapter I argue that our most concrete experiences can be conceptualized
figuratively (see also Gibbs’ most recent work presented at the RaAM 12 con-
ference in Hongkong in 2018), and not only literally.1 This does not, however,
prevent such concrete concepts from being used as source domains for figura-
tive conceptualization. We seem to have a huge body of figuratively concep-
tualized concrete experiences that we take over from previous generations in
the form of concepts, use them as if they were exclusively literally construed
experiences (concepts), and turn them into vehicles for the further conceptual-
ization of more abstract entities. Much of what we take to be literal is only
ontologically literal (i.e., constitutes basic concrete experience), but as regards
its cognitive status it is figuratively constructed experience. We reuse this
apparently concrete but figuratively understood conceptual material to com-
prehend and construct abstractions without being aware of their original
figurative (metaphorical or metonymic) cognitive status.
The fact that even basic concrete experience is commonly construed figura-
tively, that is, by means of metaphor and metonymy, is not a new and original
observation – Nietzsche, for example, suggested that “language is a graveyard
of dead metaphors.” The notion that even our basic concrete experiences are
figuratively construed raises an important question for conceptual metaphor
theory. CMT postulates (and empirically finds) that abstractions are routinely
1
The chapter reuses materials from previously published work in a revised form. In particular:
Kövecses, Zoltán. 2017a. A radical view of the literal-figurative distinction. In A. Benedek and
Á. Veszelszki, eds., Virtual Reality – Real visuality. Virtual, Visual, Veridical. 17–28. Frankfurt
am Main: Peter Lang.
22
2.1 Delimiting the Literal 23
conceptualized as basic concrete experiences. But if this is the case and it is

also true that language, including concepts that denote basic concrete experi-
ences, is figuratively construed, why is it that we conceptualize the abstract in
terms of the concrete? In other words, our fundamental distinction between the
concrete and the abstract loses its significance if both of them turn out to be
figuratively understood. Thus, we gain nothing by conceptualizing the abstract
in terms of the concrete, so the CMT explanation for the (predominant)
unidirectionality of metaphorical mappings from concrete to abstract (and
not the other way around) loses is validity (since the concrete is also figura-
tively conceived).
In classical accounts of literal meaning (such as Searle’s, 1993), literal
meaning is identified with meaning that denotes concrete things and that is
conventionally used, including dead metaphors. This is a large and founda-
tional set into which figurative meaning can always be “translated.” The
highest level generalization that supporters of both the classical theory of
metaphor and conceptual metaphor theory make is that there is such a thing
as literal meaning. Where they differ is in the role they attribute to it. Support-
ers of the classical view claim that literal meaning is the foundation of meaning
in general and that figurative (metaphoric and metonymic) meaning depends
on it. Adherents of conceptual metaphor theory suggest, by contrast, that literal
meaning is not foundational in that sense.
But researchers in the CMT tradition greatly limit the extent of literal
meaning, that is, the kinds of linguistic expressions that can be considered to
have a literal meaning. They tend to delimit the extent of literal meaning in a
variety of ways. In the chapter, I offer a radical view of the extent of the literal,
which greatly reduces or even eliminates the literal – at least in a certain sense.
(On this issue, see also Rakova 2003.)
2.1 Delimiting the Literal

Let us begin with dead metaphors. Despite their name, dead metaphors are not
really metaphors according to the classical theory. They are expressions that
acquired a new literal meaning. As we saw above, however, conceptual
metaphor theory treats dead metaphors as metaphors, the justification being
that dead metaphors are conceptually very much “alive” as examples of fully
active conceptual metaphors. The conventionalized status of dead metaphors
does not prevent them from being active in the conceptual system via the
conceptual metaphors they belong to. (On a more refined distinction, see
Müller 2008.)
Thus, in CMT the decision concerning what is literal is made on the basis of
how an expression is processed. If it activates two conceptually linked
domains, it is metaphorical; if it does not, it is literal. Lakoff (1993: 205)
24 Abstract Figuratively, Concrete Literally, Concrete Figuratively?
handles the distinction between literal and metaphorical in the following way:
“those concepts that are not comprehended via conceptual metaphor might be
called ‘literal.’” But this still leaves us with cases that are hard to classify. For
example, it is clear that “Time passed quickly” and “The time will come” are
metaphorical due to the time is motion conceptual metaphor. But what
about early in sentences like “March comes earlier than April,” a sentence used
in one of Boroditsky’s (2001) experiments? Is early literal or metaphorical? It
is usually taken to be literal. But in the experiment Boroditsky found that
speakers of English responded faster (saying TRUE/FALSE) when they
received a horizontal prime, and speakers of Chinese were faster when they
received a vertical prime. The results of this priming study indicate that a word
we would take to be literal (since it is unrelated to motion) is comprehended
metaphorically (since speakers are affected by a conceptual metaphor – hori-
zontal versus vertical motion – in processing it). Thus, if the “comprehension
via conceptual metaphor” criterion is correct, the word would have to be taken
as metaphorical, not literal.
The criterion of “comprehension via conceptual metaphor” would give us a
similar dilemma in a large number of cases, including that of the domain of
time. As a matter of fact, the word time is open to the same question: is it literal
or metaphorical? If the word early is understood metaphorically, and thus
counts as metaphorical, then surely the word time is also understood meta-
phorically and has thus a metaphorical meaning. Or should we distinguish the
name (the word) of a domain from the domain itself (the concept of time)?
But can we separate word meaning from conceptualization in a cognitive
linguistic framework? Isn’t word meaning dependent of conceptualization
(see Langacker 1987)?
We encounter the same problem in all the cases where there is an abstract
target domain (anger, theory, time, desire, etc.) involving many lin-
guistic expressions that we would not hesitate to take to be literal (such as
anger, rage, ire, irritation, fury for anger; theory, idea, scheme, belief for
theory; time, period, moment, minute, hour for time; desire, urge, longing,
want, yearn for desire; etc.). Can abstract concepts such as these be literal?
That’s how we feel about them intuitively. But this should not be the case if we
apply the criterion of “comprehension via conceptual metaphor.” Clearly, the
concepts corresponding to them are all comprehended metaphorically via a
number of different conceptual metaphors, such as anger is fire, know-
ledge is vision, time is motion, desire is hunger, etc. My claim here
would be that abstractions can only be conceptualized metaphorically, and
consequently the linguistic items expressing these metaphorical conceptual-
izations can only be metaphorical.
However, we do not view them as such. The question is why. It seems to me
that in such cases we apply a criterion different from “comprehension via
2.1 Delimiting the Literal 25
conceptual metaphor” to establish what is literal. The criterion seems to be the

following: If a word denoting an abstraction does not involve any overt
morphological indicators or markers of more concrete source domains, the
word is metaphorical in nature. The words above do not morphologically
reflect any of the well-known conceptual metaphors (i.e., source domains)
associated with the target domains in question (anger, knowledge, time,
desire mentioned above). In this regard, they are very different from lexical
expressions denoting abstractions, such as burn (with anger), clear (idea),
passage (of time), and hungry for (love), which explicitly show the concrete
source domains to which they belong.
At the same time, the assumed literal expressions belonging to abstract
target domains such as the ones above can be regarded as complying with
the Lakoffian criterion for metaphorical meaning: “comprehension via concep-
tual metaphor,” as a result of which they would cease to be literal expressions.
They would comply with it etymologically, that is to say, in a historical
perspective. Etymologically, all of these words derive from figurative (meta-
phoric or metonymic) conceptualizations. For instance, the words rage and
fury for anger go back to Latin where they had to do with madness and the
words theory and idea in the knowledge domain have to do with vision and
seeing in Ancient Greek (see, e.g., Sweetser 1990). Both of these cases show
that the words that we take to be literal today are actually motivated by some of
the same conceptual metaphors that motivate many of today’s metaphorical
expressions; namely, anger is insanity and knowledge is vision. In
some other examples, the metaphor or the metonymy is not clear in the
etymology, but the metaphoric or metonymic origin appears to be certain.
According to etymological dictionaries (such as the Online Etymology Dic-
tionary), anger originates from the Indo-European root meaning “tight,
narrow, painfully constricted.” This could probably be considered a case of
metonymic conceptualization: the cause of anger for anger, a generic-
level metonymy for anger.
In other words, it seems that a large number of expressions (that belong to
abstract domains) are taken to be literal, but in synchrony their corresponding
concepts are metaphorically comprehended and from a historical perspective
they have metaphorical or metonymic origins. Their perception as literal seems
to be based on the fact that their morphology does not reveal to contemporary
speakers any metaphoric origins.
The same argument can be taken further. In a historical perspective, many
non-abstract concepts (i.e., concrete ones) have also emerged as a result of
figurative processes. For example, in English the word book derives from the
Proto-Germanic word *bokiz meaning “beech” according to the Online Ety-
mology Dictionary (www.etymonline.com). The explanation offered by the
dictionary is that runic writing was inscribed on beechwood tablets. We may
have a series of metonymies at work here. As another example, consider the

word storm. The same dictionary suggests that it ultimately derives from the
PIE root *(s)wer- meaning “to turn, whirl.” Again, this seems to be meto-
nymic: the effect of an event for the event. The concepts that arose
by means of certain figurative processes, such as book and storm, later
became source domains for more abstract concepts (e.g., anger is a storm).
We are of course no longer aware that these concepts were produced by
figurative means; we take them to be literal. But it is clear that the same kind
of figurative conceptual processes were responsible for their emergence as the
ones we use today.
2.2 So What Exactly Is Literal Meaning?

Most of this information is of course well-known. The reason I am discussing
the issue here is to provide an answer to the question raised in the subtitle
above. The answer seems to depend on (1) whether we allow a fuller applica-
tion of the “comprehension via conceptual metaphor” criterion in synchrony
(to cases like anger and rage) and (2) whether we allow a fuller application of
the same criterion from synchrony to diachrony. It is a further issue whether it
would be legitimate, and how legitimate it would be, to apply the principle
more extensively. My view is that if we want to approach the issue from a
coherent cognitive perspective, we can and should embrace the more extensive
versions of the principle, that is, to extend it both to all the words and
expressions that belong to a particular abstract target in synchrony and to
concrete concepts in diachrony.
In my view, the extended application of the principle would be legitimate
because, on the one hand, the linguistic expressions belonging to a particular
abstract target are all comprehended metaphorically or metonymically (like
rage and fury in anger), not just the prototypical item in the domain (i.e.,
anger for anger). And, on the other hand, the move would be legitimate
because many of our categories for concrete objects and events had been
comprehended figuratively in previous times. This latter step would also be
justified because, by taking it, we would recognize our cognitive unity with
others who lived before us. The fact that we comprehended figuratively many
concrete objects and events that we take to be literal today is a very thin line
that separates us from previous generations. Their figurative conceptualiza-
tions have become unmotivated to us since we do not know many things
about the past, and thus we do not see the morphological cues that would
reveal the figurative devices they have employed. But etymologies often tell
us these stories.
What these stories indicate very clearly is that humans have always engaged
in figurative ways of understanding the world – not only in the case of
2.2 So What Exactly Is Literal Meaning? 27
abstractions but also in comprehending the concrete, physical aspects of the

world. The main difference between the two simply is that in the case of
abstractions we have conceptually constructed previously nonexistent realms
of reality, whereas in the case of concrete things already existing entities and
events have been figuratively comprehended (such as rain as “God’s tears”).
The emphasis of conceptual metaphor theory on the former case is understand-
able: It is the realm where conceptual metaphors work most fully, and they
teach us important lessons about human cognition and the way we function in
the world. As for the latter case (i.e., in the case of existing entities), we cannot
even be certain that the same systematic mapping patterns we find in concep-
tual metaphors are applicable to the realm of the concrete. We do not know,
but paying attention to this realm may turn out to be just as informative
regarding the way we function cognitively.
The Radical Claim

Overall, then, taking the stand I sketched out here would lead to a considerable
reduction of the extent of the literal. There would remain very little that would
be understood in an entirely literal way. If we accept this position, we would
actually reverse the assumption of the classical view of metaphor according to
which it is the literal that is foundational for language and thought. Indeed, we
would go even beyond the accepted view in conceptual metaphor theory
according to which the literal is needed in order to be able to comprehend
the world and to be able to create conceptual metaphors. In the radical
alternative I suggest, the figurative would be foundational, and, in many cases,
the literal would be regarded as figuratively comprehended aspects of the
concrete realm, where we lost track of the figurative comprehension of
the concrete. At the same time, it may be that this figurative comprehension
of the concrete is not a passive one but a conceptual construction of existing
things and events. The challenge would be to see in all cases whether this
“conceptual construction” is or is not the same type of conceptual construction
that we find in creating abstractions – the well-studied area of conceptual
metaphor theory.
But all this is not to say that literal meaning does not exist. It does, and it
plays an important role in conceptual metaphor theory as well. Without literal
meaning, CMT in its present form would become incoherent. In the theory,
abstract experiences are comprehended and created via concrete ones that are
expressed by apparently literal language. If there is no literal language of some
kind (corresponding to concrete experience), no metaphorical comprehension
and creation could be seen as taking place.
What I have really argued for above is that the extent of the literal may
be considerably smaller than we commonly assume. If we do not count as
literal (1) dead metaphors, (2) abstract entities or events without explicit
morphological indicators or markers for particular source domains, and (3)
the large class of cases where a concrete entity or event is figuratively
conceptualized historically, we are left with a rather limited set of cases.
Actually, it is very difficult to assess the exact size of this remaining set (but
see, e.g., work by Borghi et al. 2017; Altarriba et al. 2004). Based on research
in a variety of disciplines, we can probably include in that set basic-level
objects and events, a variety of spatial relations, basic emotions, body parts,
and probably some others (see Lakoff 1993).
But can we? Even here, we run into problems. In all of these groups, we find
items that are comprehended/constructed figuratively. For example, the basic-
level category of wolf in Hungarian is farkas, literally meaning “the one with a
tail,” or “that which has a tail.” This is the result of comprehending wolves
metonymically due to some taboos on naming this animal in the ancient belief
system. The English preposition in corresponds to the suffix –ban/ben in
Hungarian, and the Hungarian preposition goes back to the word bél
(“bowels”) in its etymology. In this case, the body serves as the source domain
for spatial relations. The English basic emotion term anger, as we saw above,
probably has a metonymic origin. Surprise (deriving from old French, and
ultimately Latin, surprendre), another basic emotion term, is the product of a
metaphorical process that can be characterized by the metaphor surprising
(someone) is unexpected seizure/attack (see Kövecses 2015b).
Finally, the French word for the face, visage, derives from the Latin visus
“sight.” This is probably also metonymic. What the examples show is that
even concepts that we can take to be based on our most basic and direct
experiences, such as the animals that surround us in nature, fundamental
spatial relations, basic emotions, and the parts of the human body can all be
conceptualized figuratively.
In other words, under a broad definition of what is figurative, very little of the
literal remains. This could seriously undermine a major claim of conceptual
metaphor theory; namely, that a sufficiently large part of language and thought
is about concrete things and is thus literal, and that we use this language and
thought to conceptualize abstractions. What we saw above was that our most
concrete experiences can be conceptualized figuratively, and not literally. But
this does not mean that they cannot be used as source domains for figurative
conceptualization. Rather, we seem to have a huge body of figuratively con-
ceptualized concrete experiences that we take over from previous generations,
use them as if they were entirely literal experiences, and turn them into vehicles
for the further conceptualization of more abstract things. Ultimately, much of
what we take to be literal is only ontologically literal (i.e., constitutes basic
concrete experience), but as regards its cognitive status it is also figuratively
construed/constructed experience. We reuse this apparently literal conceptual
2.3 Smell as a Target Domain 29
material in a figurative way to comprehend and construct abstractions without

being aware of their original metaphorical or metonymic cognitive status.
2.3 Smell as a Target Domain

We can see the same situation in the case of the concept of smell. Probably
there is general consensus that the sense of smell is one of the most basic ways
in which humans (and animals) gather information about reality. For this reason
it serves as a perfect source domain, very much in line with what we would
expect. For instance, the concept of suspicion is conventionally conceived of as
smell, hence the conceptual metaphor suspicion is smell (as in, “I can smell
trouble”). Immorality and, more generally, badness is conceptualized as (bad)
smell (as in, “I can smell a rat”), yielding bad is smelly. And the vague and
loose concept of “general atmosphere of something” is understood in terms of
smell (as in, “What he did smells of fraud”), giving us the conceptual metaphor
the general atmosphere (of something) is smell.
Thus we assume that smell is an ideal source domain. Given this, we would
not expect it to be understood metaphorically, that is, as a target domain. But it
is conceptualized as such. How is this possible? The reason, I suggest, is that
smell, like any other concept, is characterized by a number of conceptual
dimensions, or to put it in the framework of Langacker (1987), it is part of a
domain matrix. Thus the concept of smell is connected to the idea that it is
something that exists, that it is something that can have various degrees of
intensity, that it is something that the perceiver, or experiencer, cannot help
perceiving, that is, that he or she has no control over whether or not s/he
perceives it. In other words, we can identify at least the following three
dimensions for the concept of smell: existence, intensity, and lack of control.
In a “domain matrix” view of the conceptual system, the three dimensions
are high-level, generic concepts that characterize many concepts, concrete and
abstract alike. They function as shared dimensions of concepts. Existence,
intensity, and (lack of ) control are concepts that form an integral part of many
domains, including smell. An important property of these dimensional con-
cepts is that they are all generic-level ones and abstract. Such generic-level
abstract concepts account for the metaphoricity of even the most tangible
concepts, such as smell. The question is what kinds of conceptual metaphors
are formed by making use of them in relation to smell. (The remainder of this
section uses materials from Kövecses 2017a.)
Existence
Let us begin with existence. Consider some examples for how the existence of
smell is presented by a dictionary:
VERB + SMELL be filled with, have The air was filled with a pervasive smell of
chemicals. The cottage had a musty smell after being shut up over the winter. | give off
The skunk gives off an unpleasant smell when attacked. | catch, detect As she walked
into the house she detected the smell of gas. (boldface in the original)
The examples indicate a variety of ways in which the existence of smell can be
conceptualized. In the case of fill, the smell is viewed as a substance and the
object or location with the smell is viewed as a container, and the existence of
the smell as the substance being in the container. In the case of have, existence
of the smell corresponds to the possession of an object that is smell itself. In
the case of give off, causing smell to come into existence is the transfer of an
object and the source of the smell is a surface. Finally, catch and detect are not
dedicated to denoting the existence of smell, they simply assume its existence.
The examples and their construals of the existence of smell they reveal point
to the following conceptual metaphors:
filled with:
smell is a substance
the object / location that has the smell is a container
the existence of smell is for the substance to be in the container
have:
smell is an object
the existence of smell is the possession of an object
give off:
smell is an object
causing smell is transferring the object
source of smell is a surface
The three cases conceptualize the experience of smell as the existence of a

state. The state is either a substance or an object. Existence is viewed as a
substance being in a container, as possessing an object, and as transferring an
object. The three specific cases of existence metaphors for smell are thus
instances of three generic-level metaphors for existence:
existence is being in a container
existence is possession
(causing) existence is transfer
These conceptual metaphors apply to a large number of specific-level instances
of existence, not just to the existence of smell.
Intensity
Another generic-level concept that is linked to smell is intensity. In the case of
smell, intensity involves a scale of values from weak to strong. Collocation
dictionaries give us a good idea of this notion in relation to smell:
2.3 Smell as a Target Domain 31
ADJ. overpowering, pervasive, pungent, rich, sharp, strong There was

an overpowering smell of burning tyres. | faint | distinct | distinctive,
particular, unmistakable | funny, peculiar, strange, unusual What’s that
funny smell? | familiar | lingering | aromatic, delectable, delicious, fra-
grant, fresh, lovely, nice, savoury, sweet, wonderful the aromatic smells
of a spring garden full of herbs | warm | appalling, awful, bad, evil,
horrible, nasty, offensive, terrible, unpleasant, vile | acrid, nauseating,
putrid, rank, sickly An acrid smell filled the air. | damp, dank, musty,
rancid, sour, stale the sour smell of unwashed linen | earthy, fishy,
masculine, metallic, musky, oily, smoky, spicy | cooking Cooking smells
drifted up from the kitchen. (boldface in the original)
Many of these adjectives deal with the different kinds of smell, but some of
them denote the intensity aspect of smell. These latter adjectives (and the verb
fill) appear to be based on several different conceptual metaphors – in some
cases, on etymological grounds:
overpowering, strong, faint: intensity of smell is strength of effect
sharp, pungent: intensity of smell is sharpness of an object
vaguely: intensity of smell is degree of brightness of light
pervasive, fill: intensity of smell is quantity of a substance
These conceptual metaphors that relate to the intensity aspect of smell also
have their generic-level variants:
overpowering, strong, faint: intensity is strength
sharp, pungent: intensity is sharpness
vaguely: intensity is brightness
pervasive, fill: intensity is quantity
Lack of Control
The issue of control is also relevant to smell and perception in general.
A curious property of perceptual experience is that we do not have control
over what we perceive. If our sense organs are normal and no external
circumstance blocks the perception, we undergo the perception of any existing
perceptual stimulus. This is also often rendered in metaphorical ways – one of
the chief ways among them (according to dictionary information) is by means
of the verb hit in the case of smell. Consider the example taken from a
dictionary: Then the pungent smell hit us - rotting fish and seaweed.
We can analyze the metaphorical use of hit as follows:
smell is a physical force
intense smell is a strong physical force
sensing an intense smell is coming into contact with a
strong physical force
The three basic conceptual metaphors above can be combined into a complex
one that describes the metaphorical conceptualization of the perceptual experi-
ence of strong smell:
the experiencer of an intense smell is a person coming into
abrupt physical contact with a strong physical force
But most importantly, the main point is, as before, that the specific-level
conceptual metaphors are licensed by corresponding generic-level ones:
causes are forces
intensity is strength
sensing is contact
Again, these generic-level metaphors pervade the conceptual system and allow
us to conceptualize various kinds of causes, intensity, and sensation, not just
smell as a cause, not just the intensity of smell, and not just sensing smell. In
other words, smell as an extremely basic and direct experience has aspects to it
that lend themselves to metaphorical conceptualization.
2.4 Conclusions
On the basis of the evidence we saw in this chapter, it seems that figurativity
(i.e., understanding concepts by means of metaphor or metonymy) does not
really distinguish concrete concepts from abstract ones. Both types of concepts
(concrete and abstract) are characterized by being understood metaphorically
and metonymically. As a result, we face a dilemma in accounting for the
unidirectionality of metaphorical mappings in conceptual metaphors. We
cannot claim that the concrete is understood in a literal way and we make
use of this in comprehending and creating abstractions.
Does this situation mean we have to discard CMT? In my view, it should not
mean that. We can propose, given the evidence we saw, that concepts consist
of two parts: an ontological part that specifies what kind of ontological
materials make up a concept and a cognitive part that specifies the ways in
which concepts are cognitively construed or constructed. I assume that this
distinction is similar, but at least related, to a distinction made by Langacker
(1987, 2008) between the content part and the construal part of concepts.
Now, given the distinction, we can suggest that in the case of literal meaning
for concrete concepts, the ontological content predominates over the cognitive-
construal part. In many instances, the construal part is limited to certain
generic-level dimensions of concepts, as we saw in the case of smell. These
dimensions that are shared across many concepts in the conceptual system are
figuratively (metaphorically, metonymically) understood. By contrast, in the
case of abstractions, the construal part predominates over the ontological part.
2.4 Conclusions 33
The cognitive-construal part of abstract concepts is essentially figurative

(metaphoric, metonymic). Their actual ontological part (i.e., their ontological
content) is small. In experientalist philosophy (see Lakoff and Johnson 1999),
we call this small ontological essence or core “embodiment.” Embodiment
retains the tangible content ontology of abstract concepts. This is what cogni-
tive linguists and scientists mean when they talk about abstract concepts as not
being “transcendental” but embodied.
Thus, both concrete and abstract concepts have both embodied content
ontology and figurative construal (i.e., figuratively constructed understand-
ing) – but in different proportions. In conceptual metaphors, we have predom-
inantly content-ontology-based concepts as source domains and predominantly
figuratively-construed concepts as target domains. There are probably no pure
ontology-content-based concepts and no pure figuratively-construed concepts.
This view explains, in addition, why an “ideal” source domain, such as smell,
can also function as a target domain. The reason is that we can profile the
ontology part in some cases (and we get a concept as a source domain, such as
smell as source), but in others we can profile the figuratively construed/
constructed part (and we get target domains, such as smell as target).
3 Direct or Indirect Emergence?
Primary metaphors form the foundation of conceptual metaphor theory.1 They

are foundational, in that they are seen as directly emerging from our most basic
embodied experiences and, also, in that they constitute complex metaphors.
I will ask: Do they really emerge directly or do they emerge through a
metonymic stage? There is a debate between scholars who suggest that many
metaphors are based on, or derive from, metonymies, versus those who do not
see such connection between the two. “Resemblance metaphors” do not seem
to have anything to do with metonymy. However, in the case of “correlation
metaphors” (on these, see, e.g., Grady 1997a,b, 1999; Lakoff and Johnson
1980, 1999), several researchers argue that metaphors arise from, and are not
independent of, metonymies.
What is the relationship between metaphor and metonymy in our conceptual
system? This issue is perhaps as old as the study of metaphor and metonymy,
but I do not wish to approach it from a historical perspective. The issue that
I address here is whether metonymies, or, more precisely, metonymic thinking,
play any role in the emergence of metaphors. My starting point will be the
debate taking place in especially cognitive linguistic circles between scholars
who suggest that many metaphors are based on, or derive from, metonymies,
as opposed to those who do not see such connection between the two.
The issue is relatively uncontroversial in the case of what are called “resem-
blance metaphors” (see Grady 1999), which do not seem to have anything to
do with metonymy (unless we take the partial structuring of a target domain by
a source domain metonymic in character, where the inevitably partial structure
of the source, B, is used to conceptualize an equally inevitable part of the
target, A, resulting in the metonymies a part of b for the whole of b
and a part of a for the whole of a, given the general metaphor format
a is b. For this view, see, especially, Barcelona 2000). However, in the case of
another kind of metaphor, called “correlation metaphors” in the cognitive
linguistic literature (see, e.g., Grady, 1997a,b, 1999; Kövecses, 2002/2010;
1
This chapter is a revised and enlarged version of Kövecses (2013).
34
3.1 A Brief Note on the Conceptual System 35
Lakoff and Johnson 1980, 1999), metonymy appears to be more clearly and
importantly connected with metaphor, as argued by several researchers, such
as Barcelona (2000); Kövecses (2002/2010); Kövecses and Radden (1998);
Lakoff and Kövecses (1987); Radden (2002); Radden and Kövecses (1999);
Taylor (1989), and others (in, e.g., the volume edited by Dirven and Pörings
2002). In contrast, several recent influential researchers within the conceptual
metaphor paradigm believe that correlation metaphors (best known as
“primary metaphors”) come into existence independently of, or without any
recourse to, metonymy (most notably, Grady 1997a,b; Grady and Johnson
2002; Lakoff and Johnson 1999).
As will be clear from the chapter, I align myself with the former group and
suggest that many metaphors (of the correlational kind) derive from metonymies,
that is, they have a metonymic basis. What distinguishes my position from the
view of the other proponents in the group that favors a metonymy-based
emergence for many metaphors is that I attempt to establish the relationship
between metaphor and metonymy by relying on several particular characteristics
of the conceptual system, as we know it today.
My specific proposal in the chapter will be that correlation-based metaphors
emerge from frame-like mental representations through a metonymic stage.
I suggest this happens when one of the elements of a frame-like mental
structure is generalized (schematized) to a concept that lies outside the initial
frame in a different part of the conceptual system. The generalization process
leads to sufficient conceptual distance between the initial and the new frame on
which metaphors can be based.
3.1 A Brief Note on the Conceptual System

I take the conceptual system to be a structured organization of concepts and a
set of cognitive operations that are used to make sense of the world. In the
chapter, I make especial use of certain aspects and certain cognitive operations
of the system. In particular, I assume that the conceptual system is hierarchic-
ally organized (see, e.g., Rosch 1978), consists of frame-based concepts (see,
e.g., Barsalou 1999; Fillmore 1982; Lakoff 1987; Langacker 1987), is dynamic
in its functioning (see, e.g., Barsalou 1999; Gibbs 2003; Gibbs and Cameron
2007; Langacker 1987), and is embodied in nature (see, e.g., Gibbs 2006;
Johnson 1987; Lakoff 1987). In addition to focusing on the cognitive oper-
ations of metaphoric (see, e.g., Gibbs 1994; Kövecses 2002/2010; Lakoff and
Johnson 1980, 1999) and metonymic conceptualization (see, e.g., Barcelona,
Benczes, and Ruiz de Mendoza 2012), I will also rely on two further cognitive
operations in this chapter – generalization, or schematization, and specializa-
tion, or elaboration, as these have come to be used in cognitive linguistics
(see, especially, Langacker 1987, 2008).
There are two kinds of structure that characterize the conceptual system.
Conceptual systems are organized “vertically,” which, essentially, provides for
a thematic structure in the system (see, e.g., Rosch 1978) and also “horizon-
tally,” which, essentially, defines individual concepts and consists of smaller
domains, or frames, or idealized cognitive models (see, e.g., Barsalou 1999;
Fillmore 1982; Lakoff 1987; Langacker 1987).
3.2 Metaphor or Metonymy?

There are many cases where it is difficult to distinguish metaphor from
metonymy, given “standard” definitions of metaphor and metonymy (see,
e.g., Goosens 1990). A common definition of metaphor in conceptual meta-
phor theory is that in metaphor we conceptualize one domain in terms of
another (Lakoff and Johnson 1980). In metonymy, an element in a domain, or
frame, provides mental access to another element within the same domain,
or frame (Kövecses and Radden 1998; Radden and Kövecses 1999). The cases
where metaphor and metonymy are difficult to distinguish are those where it is
not clear whether we deal with one domain, or frame, or two (see also, Croft
1993). My suggestion is that in order to be able to resolve the dilemma, we
have to take into account the larger structure of the conceptual system; namely,
the structure consisting of both thematic hierarchies and frames, as well as the
cognitive operations of generalization (schematization) and specialization
(elaboration). Broadly speaking, two cases can be distinguished. In the first,
there is an initial frame and an element of this frame is generalized to become
the source concept of a conceptual metaphor. In the second, there is a frame
and an element of this frame is generalized to become the target concept of a
conceptual metaphor.
Case 1: A Frame Element Becomes the Source

In this case, a frame that becomes the metaphoric target domain involves an
element that becomes the metaphoric source domain: This element is general-
ized to a concept outside the original frame. For an initial illustration of the
difficulty concerning whether particular expressions are metaphoric or meto-
nymic, consider as an example some emotion-related expressions:
He is in low spirits. (sadness)
She is feeling up. (happiness)
He is a hothead. (anger)
Are these metaphoric or metonymic expressions? We could argue both that
they are metonymic and metaphoric. To be physically down is a part of our
folk theory of sadness, to be upward-oriented is a part of our conception of
3.2 Metaphor or Metonymy? 37
happiness, and an increase in body heat is an important ingredient of our

everyday idea of anger. And, of course, our folk theories of sadness, happiness,
and anger each constitute single coherent domains, or frames, mentally repre-
senting the concepts, in which we have the elements of being physically
downward-oriented, upward-oriented, or characterized by a higher than normal
body temperature, respectively, for sadness, happiness, and anger. In other
words, in all these cases we have a single domain, or frame, for the concepts
of sadness, happiness, and anger, where an element of the frame is used
for the whole frame; that is, we have to do with metonymies, given our
definition above.
But we know that, for example, in the 1999 Lakoff and Johnson view as
presented in their Philosophy in the Flesh, these are some of the prime
examples of conceptual metaphors. Following Lakoff and Johnson, we would
not hesitate to assign these linguistic expressions to the conceptual metaphors
sad is down, happy is up, and anger is heat, respectively. So what is
the appropriate way of thinking about such cases? Should we claim that the
expressions are based on the conceptual metonymies the downward
orientation of the body for sadness, the upward orientation
of the body for happiness, and an increase in body temperature
for anger? Or should we suggest, instead, that they are based on the
conceptual metaphors sad is down, happy is up, and anger is heat?
Let me offer the following solution to the dilemma: We have certain
behavioral responses associated with these emotions. For example, sadness
includes drooping body posture, mouth turned down, etc. These are general-
ized into the concept of “downward bodily orientation.” Given the general
effect for cause metonymy (on schematic metonymies, see Ruiz de
Mendoza and Mairal 2007) and given that behavioral responses function as
metonymies in emotion concepts (Kövecses 1986, 1990, 2000b, 2008; Lakoff
and Kövecses 1987), we get the specific metonymy downward bodily
orientation for sadness. All of this happens inside the sadness frame,
since downward bodily orientation is one of the behavioral responses
associated with sadness.
But downward bodily orientation can also lead to metaphoric conceptual-
ization in two, possibly successive, ways. The first is that “downward bodily
orientation” can be generalized to the concept of down(ward). This (i.e.,
down) is a spatial concept that, in virtue of the process of generalization (or
schematization) from behavioral responses characterized by a downward
spatial orientation in sadness, is divorced and distinct from the actual behav-
ioral responses associated with sadness inside the sadness frame. The gener-
alized, or schematized, concept of down(ward) is in the space thematic
hierarchy, and not in the emotion one, where the more specific responses
characterized by a downward orientation can be found. And since it is divorced
and distinct from them (i.e., is outside the sadness frame), it can now be seen
as a (source domain of a) metaphor for sadness: hence the conceptual metaphor
sad is down. sadness is in the emotion-dominated part (target) of the
hierarchy, whereas down(ward) is in the (more concrete) space-dominated
part (source).
The conceptual metaphor sad is down resulting from the generalization,
or schematization, of various behavioral responses motivates, or, to use
different cognitive linguistic terminology, licenses, or sanctions, a number
of metaphorical linguistic expressions, such as “to feel down,” “to be in low
spirits”, and the like. But, possibly based on this generalization, we can
develop further metaphorical conceptualizations in the following way: The
notion of down can be elaborated by a large number of more specific
instances of space that have nothing to do with sadness as conceived in the
sadness frame, including places with a downward orientation, such as pits
(resulting in “be in the pits”), dumps (resulting in “be down in the dumps”),
and the like. This is the second way in which downward bodily orientation can
lead to metaphoric conceptualization.
Thus, given the example above, we can suggest that metaphor appears to be
based on and derive from metonymy (as suggested by Lakoff and Kövecses for
anger and by several other authors, e.g., Barcelona 2000; Radden 2002). The
details of this development in many cases, such as the sad is down metaphor,
involve the process of generalization, or schematization, from specific bodily
responses characterized by a downward orientation (leading to a more general
spatial concept, down, in the case of sadness) and (then) the process of
specialization, or elaboration (leading to elaborations of the same schematic
spatial concept). Kövecses and Radden (1998) make a similar argument for the
source domain of heat in relation to anger in the conceptual metaphor anger
is heat (Lakoff and Kövecses 1987).
Consider now another example – the relationship that obtains between
closeness and intimacy. This is usually conceptualized, and viewed, as the
metaphor intimacy is closeness. But, similar to the previous cases, we can
also conceptualize the relationship between the two concepts as metonymy.
Since children and adults in general who are in an intimate relationship with
each other tend to be physically close (and, as a result, be able to touch each
other, feel the body temperature of the other, be able to smell the other, etc.),
we can think of it as a metonymy: closeness for intimacy. However, the
physical closeness that characterizes intimacy is specifically between two
people – with the several specific consequences just mentioned. Thus the less
generic version of the metonymy is physical closeness between two
people for intimacy. Similar to the generalization process we have seen
for sadness above, the physical closeness between two people is
then generalized, or schematized, to the concept of closeness. The words

“close,” “distant,” “remote,” and others, describe closeness (or a lack of it)
between any two physical objects, not just people. Since physical close-
ness between two people is a conceptual element of a single coherent
functional domain, the intimacy frame, we can think of the relationship
between this element and intimacy as a metonymic one.
At the same time, the more general concept of closeness is part of a
taxonomic, or thematic, hierarchy – the “vertical” domain of space. When this
more general concept is used in the mapping, we get the metaphor intimacy
is closeness. In addition, the generalized, or schematized, concept of
closeness can undergo specialization (elaboration). When this happens, we
get more specific linguistic expressions to talk about intimacy (or a lack of it),
including “bond,” “tie,” “tight as a glove,” and “remote.” It is clear that “bond”
and “tie” represent instances of closeness, in that they are manifestations of
two objects being close. “Remote” is special because it primarily obtains
between physical places – not physical objects. In other words, the relationship
between closeness and intimacy can be analyzed in the same way as that
between down and sadness.
The above analysis leads to the conclusion that, in at least some cases,
metaphors derive from metonymies through the application of the cognitive
processes of generalization (schematization) and specialization (elaboration).
The way this can happen is that there is a particular frame with a specific
element inside the frame, and the element can be used to provide access to the
whole frame, i.e., it can be used metonymically. This is the metonymic stage of
the process. Further, the element inside the frame involves or is characterized
by a particular frame-specific concept, which is, then, generalized, or schema-
tized, into a concept that exists outside the frame (often in a different taxo-
nomic hierarchy). When this happens, we have to do with metaphor, i.e.,
where the initial frame, or domain, is conceptualized in terms of another,
conceptually distant frame, or domain (where conceptual “distance” derives
from being in a different taxonomic hierarchy).
It could be asked how general this account actually is. Can it be that it only
applies to emotion metaphors and metonymies? To see this, let us now take
another metaphor that comes from outside the emotion domain. Consider now
control is seeing (see Sweetser 1990), as exemplified by expressions like
“see to something” (“I’ll see to it . . .”), “keep an eye on something” (“I’ll keep
an eye on it for you.”). It can be suggested that we have a frame (or scene) in
which visual monitoring is one of the things we do when we guard or otherwise
keep control over an entity. In the initial frame, there is a correlation between
visual monitoring and guarding. Since the correlation occurs within the same
frame (or scene), we can use one element of the frame (visual monitoring) to
have access to the whole frame (guarding) – a situation that yields the meton-
ymy visual monitoring for guarding (or other kinds of control).
Visual monitoring is a kind of seeing, just as guarding and other forms of
keeping control are kinds of control. Given the initial frame in which they
are correlated, when visual monitoring is generalized (schematized) as seeing
and when various forms of control, such as guarding, are generalized
(schematized) as the notion of control in very different vertical hierarchies
of the conceptual system (control vs. vision), we conceptualize control
through seeing. We can think of this relationship as a metaphor, rather
than metonymy, because seeing and control are in distant parts of the
conceptual system (perception vs. social relation, or the like). But, in all
probability, they are brought together in the metaphor in the first place because
their more specific versions (visual monitoring and guarding) are both parts of
the same (initial) frame, or functional domain (or scene). Without this
co-occurrence in the same frame (yielding a metonymy), we would probably
not have thought of conceptualizing the control domain as seeing, yielding
a metaphor.
Given this analysis, the following general picture seems to emerge: One of
the events or states within a single initial frame (the original frame that
becomes the metaphorical target domain, such as intimacy or guarding/
control) gives rise to a distinct frame that becomes the metaphorical source.
We can represent this as follows:
ELEMENT 1 metonymy ELEMENT 2

schematization
(PROSPECTIVE
TARGET) SOURCE
(PROSPECTIVE
SOURCE)
metaphor
Figure 3.1 Schematization of the prospective source domain

In the examples, we have seen so far, there is a correlation between two
events or states within the same frame, or scene (e.g., physical closeness
between two people and being in an intimate relationship within the intim-
acy frame or visual monitoring and guarding within the guarding frame),
such that one of the events or states (e.g., being in an intimate relationship or
guarding someone or something) is accompanied by the other event or state
(e.g., being physically close to another person and visually monitoring some-
one or something) that is generalized, or schematized (e.g., as closeness and
seeing). In other words, in such cases the original frame that becomes the
metaphorical target gives rise to the source in the metaphor.
Case 2: A Frame Element Becomes the Target

In this case, a frame that becomes the metaphoric source involves an element
that becomes the metaphoric target: This element is generalized to a concept
outside the original frame. Unlike Case 1, in another set of well-known
metaphors, we find the opposite process: The original frame that becomes
the metaphorical source gives rise to the metaphorical target. There is a
correlation between two events or states within the same frame, or scene, such
that one of the events or states gives rise to the event or state that becomes the
target domain. In such cases, one of the events or states in the original frame,
or scene, that becomes the source domain of the metaphor gives rise to the
target. One such metaphor is knowing/ understanding is seeing.
When we see something, it enters our awareness, and we can examine it,
identify its size, weight, color, etc., we can describe it; thus, we can find out about
it, we can know it. Seeing something physical commonly makes knowledge about
things possible and knowing something commonly requires seeing things. The
relationship between seeing something physical and knowing it can be thought of
as a correlation between two experiences (seeing and knowing) within the same
frame-like structure, and we can think of one event (seeing) as enabling the other
(knowing). When seeing something physical and knowing the thing are co-present
and correlated in the same frame-like mental structure, we can regard their
relationship as a metonymic one and use one for the other, where the metonymy
would be seeing something physical for knowing the thing.
However, not all forms of knowledge involve physical things that can be
known by seeing. Our knowledge regarding nonphysical things, such as ideas,
theories, facts, feelings, memories, are of an entirely different sort. We do not
know and understand them because we physically see them, but we can
conceptualize how we know or understand nonphysical things on the basis
of what happens when we see something physical and know it. This requires
that we generalize the concept of knowledge as based on physical things to a
concept of knowledge that is independent of our perceptual experience.
Knowledge thus becomes a concept (representable by means of a frame-like
mental structure) that exists independently of our perception-based experi-
ences (such as seeing). This highly schematic concept can become the target
domain, and we can conceptualize knowing something nonphysical as seeing
something physical; hence the metaphor knowing is seeing. (A similar
argument, following and reassessing Traugott and Dasher 2002, for the devel-
opment of this metaphor can be found in Sullivan 2006.)
Just as in the previous cases (Case 1 above), the question arises whether the
course of development outlined above for the knowing is seeing metaphor
is a unique feature of this metaphor or it applies to other metaphors as well. As
a further example, let us take the achieving a purpose is arriving at a
destination (better known as purposes are destinations) metaphor.

We can suggest that here as well we have a frame-like mental structure, in
which there are two correlated events: arriving at a destination and achieving a
purpose. In those cases where our only goal is to get to a destination, arriving
at the destination implies achieving our goal. Thus there is a fundamental and
inherent correlation between destination and purpose that characterizes the
“going to a destination” frame. But in other cases, we want to reach a
destination because we have additional purposes to achieve at those destin-
ations. These two types of situation indicate that, often, reaching a destination
is necessary to achieve our goals (either getting to the destination or getting to
the destination to do something else there). Thus, the relationship between the
two events can be taken to be based on a conceptual metonymy, where an
element of the frame can be used for the other element, yielding the metonymy
arriving at a destination for achieving a goal.
This kind of direct goal achievement can be extended to cover still other
cases, where we have goals that have nothing to do with going to a destination.
There exists an extended and general meaning of the concept of purpose that is
entirely independent of going to a destination. We may have goals that have
nothing to do with going to destinations. Thus, the purpose is outside the “going
to a physical destination” frame; it is distinct from it and is in a distant taxonomic
hierarchy. This is a highly generic and schematic idea of purpose – the kind of
purpose that can be conceptualized metaphorically by means of the “going to a
physical destination” frame, or scene. This way, the “physical destination kind
of purpose” leads to a highly schematic concept that becomes the target domain
of the metaphor achieving a(ny kind of ) purpose is arriving at a
destination (or purposes are destinations, for short).
In other words, in the kinds of cases we have seen in Case 2, given a frame, an
event or state (e.g., seeing) that becomes the source domain of a metaphor gives
rise to an event or state (e.g., knowing) that becomes the target. This is the opposite
of the previous cases discussed under Case 1, where, given a frame, an event or
state that becomes the target domain gives rise to an event or state that becomes the
metaphoric source. Case 2 can be given visual representation as follows:
ELEMENT 1 metonymy ELEMENT 2

schematization
(PROSPECTIVE
TARGET)
TARGET (PROSPECTIVE
SOURCE)
metaphor
Figure 3.2 Schematization of the prospective target

3.3 Some Further Issues 43
However, in both cases, given two events or states in an initial frame, one
event or state stands metonymically for the other event or state within the same
frame, and, then, one of the events or states is generalized (schematized) in
such a way that it becomes independent of the original frame. When the latter
happens, we find the two initially co-occurring and correlated events or states
within the same frame emerging in two distinct and conceptually distant
frames in different parts of the conceptual system. We can now think of such
cases as metaphoric connections between the frames.
3.3 Some Further Issues

In this section, I take up three issues that are relevant to the discussion of the
metaphor-metonymy relationship. They are (1) the question of whether there
are similar approaches to the problem, (2) the issue of the components of
scenes, and (3) the nature of correlations.
Similar Approaches
The idea that an event or state can stand metonymically for another event or
state within the same frame, or frame-like mental structure, and that one of the
events or states can be generalized, or schematized, to a concept that exists
independently of the original frame (thus leading to a metaphor) is compatible
with several approaches that provide accounts of how metaphors and meta-
phorical meanings emerge.
One such approach is Grady and Johnson’s (2002) work, where they note that
some metaphorical usages are associated with scenes where the literal and
metaphorical meanings of words (like the visual and cognitive senses of “see”)
are not initially distinguished by children; the two meanings of a word are
conflated. Later on, however, the meanings are distinguished by children, that
is, deconflated. The notions of conflation and deconflation are Grady and
Johnson’s way of accounting for the emergence of primary metaphors.
A primary scene, such as that of “becoming aware by seeing,” consists of two
subscenes: in this case, the perceptual subscene of seeing something and the
cognitive subscene of becoming aware of something. For example, in the process
of the emergence of the metaphorical sense of “see,” the child learns to differen-
tiate the cognitive subscene and the perceptual one, which gives rise to the
primary metaphor knowing/ understanding is seeing. In the account
I am proposing, the stage of deconflation could be seen as being preceded by a
metonymic stage, where one of the events or states stands metonymically for the
other within the same primary scene. Furthermore, I would also suggest that
deconflation can occur after the generalization, or schematization, of one of the
events or states has taken place. This stage can be seen as opening the way to the
emergence of metaphors – the stage where an event or state is generalized to a
concept that exists independently of the scene, or frame-like mental structure, in

a distinct and distant part of the conceptual system.
We can find some evidence for this account in some corpus-linguistic
research. In her argument against a rigid distinction between literal and
metaphorical meaning Deignan (2005: 66) presents a set of linguistic examples
that seem to indicate the three stages of the sense development mentioned
above in connection with the verb see, as it is used in the controlling is
seeing conceptual metaphor. Consider the following examples of the use of
the idiom keep an eye on:
(1) This means that while Julia is cooking she can still keep an eye on their
two young children.
(2) [The job] entails collecting the rent and keeping an eye on some housing
association flats.
(3) Often a scheme will need backing for several years. During this time the
Field Director [will] keep an eye on its progress.
The idiom keep an eye on has to do with the conceptual metaphor control-
ling is seeing. Deignan argues that in the first example keep an eye on is
metonymic, in the second it is ambiguous between a metaphoric and meto-
nymic interpretation, and in the third it is metaphoric. Example (2) that is
ambiguous between a metaphor and a metonymy corresponds to the deconfla-
tion stage and example (1) to the metonymic stage in my interpretation. It
should be noted, however, that this corpus-linguistic evidence is only partial. It
relates to the contemporary simultaneous existence of the metonymic,
metonymic-metaphoric, or metaphoric uses of the expression. It does not tell
us anything about the sequence in which this differentiation may have evolved
through time.
Several other authors make a more or less similar argument concerning the
metaphor-metonymy relationship, but they all use conceptual machinery that is
different from the one I use here (see, e.g., Barcelona 2000; Taylor 1989; and
several papers in Dirven and Pörings 2002). For example, a piece of work that
seems fully compatible with the approach advocated here is Radden’s (2002)
paper. Radden argues that in a large number of cases metaphors derive from
metonymies and that they emerge through a continuum involving literal
meaning through metonymy to metaphor. He distinguishes four cases where
metaphorical meaning, and more generally, metaphors, can derive from meton-
ymy: (1) correlations in experience, (2) conversational implicature, (3) taxo-
nomic structure of categories, and (4) cultural models. The obvious similarity
between his approach and mine is that we both see many metaphors as
deriving from metonymy. As a matter of fact, several of the examples dis-
cussed in his paper are the same as the ones discussed here. An important
difference, though, between his proposal and mine is that while Radden
identifies four different sources for the emergence of metaphors from meton-
ymy, I offer a more unified view that crosscuts his four sources and that
accounts for metonymy-based metaphors in terms of two events or states being
correlated in experience, a situation that is represented conceptually by means
of a frame-like mental structure. In other words, I view correlations between
two events or states as involving or underlying three of his four proposed
metonymic sources. (Radden also recognizes correlation as a metaphor-
producing source in his first group of sources.)
The view of the metonymy-metaphor relationship as based on correlations
between elements of a frame would bring the account I propose closer to
Grady’s notion of primary metaphor that is also based on the notion of
correlation. However, Grady’s view and mine are different on a number of
points. Grady and Johnson (2002: 540) define primary metaphors as follows:
“. . . primary metaphors are motivated by tight correlations between distin-
guishable dimensions of recurring, locally defined experience types. We refer
to these dimensions, which unfold dynamically over very brief time spans, as
subscenes.”
And in the same paragraph they continue: “Note that because subscenes are
co-occurring aspects of simple scenarios, our account might be taken to
suggest that primary metaphors arise from metonymies; it is important to
consider, however, that a metonymic relationship concerns conceptual and
referential association, whereas our proposal refers to correlations at the level
of experience, and to truly metaphoric patterns of conceptualisation which
arise from these correlations.”
Here Grady and Johnson explicitly deny the possibility that primary meta-
phors arise from metonymies – the claim I have made. They base their
argument on the idea that while metonymy “concerns conceptual and referen-
tial association,” primary metaphors concern correlations in experience. But in
order to become linguistically coded, experiential correlations must be con-
ceptualized, and once conceptualized, experiential correlations also become
conceptual (associations). Grady and Johnson also argue that primary scenes
and the subscenes that constitute them have a unique status among the
conceptual tools (such as domains, frames, mental spaces, schemas) used to
represent categories in the conceptual system. If valid, this argument could be
used to suggest that since metonymies can be found in frames, and since
primary scenes are different from frames (and other similar conceptual
structures), primary metaphors cannot derive from metonymies (which are tied
to frames). However, it is notoriously difficult to distinguish frames, mental
spaces, domains, schemas, scenarios, scripts, and, importantly, scenes from
each other in such a way that a distinction between primary metaphor and
metonymy could be based on the distinction. All of these conceptual structures
are mental representations of coherent aspects of human experience. As such,
despite their various differences in abstractness and the kinds of aspects of the
world they are taken to represent in the various disciplines (see, e.g., Andor
1985), they are all coherent mental representations of aspects of the world
consisting of parts and being conceptualized as wholes. (This is why in several
places in the chapter I have used the term “frame-like mental structure” that fits
the general, overarching definition. In the next chapter, Chapter 4, I will take
up this issue for a different purpose.) The correlations between events, states,
actions, etc. (the subscenes) that make up primary scenes may be thought of as
coherent organizations of human experience (i.e., frame-like conceptual struc-
tures) as well, with parts constituting wholes. Because of this, metonymies
may be based on primary scenes as well, and, as proposed above, may in turn
lead to metaphors (for an early formulation of this idea, see, e.g., Kövecses,
2002/2010; Kövecses and Radden 1998; Radden and Kövecses 1999).
In a more recent paper, Grady (2005) is careful to point out that many
metonymies are also correlations between concepts (such as author for
work, place for institution, means for action), or to put it differ-
ently, some correlations lead not to metaphoric connections but to metonymic
ones. Grady proposes that those correlations in experience lead to primary
metaphors that meet three conditions: (1) the correlation is between a sensory
and a nonsensory (mental) experience; (2) the two correlated experiences share
highly schematic structure; and (3) the correlated experiences covary, that is, if
there is a change in one, there is a similar change in the other.
In other words, on this view, a small group of correlations (characterized
by the three properties) gives rise to metaphor, not to metonymy. As an
alternative account, I would propose to consider this special correlation type
as deriving from metonymy because, as I argued, the components of primary
scenes function as parts (elements) within a single frame-like mental struc-
ture conceptualized as a whole, where a component part can provide access
to either another component part or to a whole (and then, through
generalization, paving the way for the emergence of full-blown cases of
metaphor). As a matter of fact, the sensory-nonsensory condition that char-
acterizes metaphors in the theory of primary metaphor may well be thought
of an equally good condition for metonymy. Given a frame or (primary)
scene, the conceptual element that represents something sensory in the frame
or scene is typically selected to stand for something nonsensory. That is, the
sensory-nonsensory distinction can be regarded as a further member of the
series of factors (or conditions), such as “human over nonhuman,” “visible
over nonvisible,” “specific over nonspecific,” etc., that are commonly evoked
to explain which element can provide access to another element in the most
typical manifestations of metonymy (Kövecses and Radden 1998; Radden
and Kövecses 1999).
Components of Scenes
I suggested above that scenes are like other frame-like mental structures that
can lead to metonymy (and then metaphor). The question I would like to take
up now is what the components are, or can be, that make up such frame-like
structures. One way in which primary scenes are distinguished from other
frame-like structures is that they consist of two subscenes of events, actions, or
states, both involving at least one entity and one relation (e.g., x sees y and x
knows y). These events, states, and the like (the subscenes), are then taken to
be correlated and form the basis of primary metaphors. But can we distinguish
primary scenes from other frame-like representations on this basis? Is it the
case that frame-like mental structures that can give rise to metonymy consist of
entities (representing individuals and relations) only and that they cannot
include (pairs of ) combinations of entities and relations (i.e., subscenes) of
the kind we have seen, for example, for knowing is seeing? This is not the
case. Recent metonymy researchers commonly work with notions like com-
plex events and subdomains in their studies of metonymy, which often contain
combinations of entities and relations (similar to “x sees y”). In other words, it
appears that the proposed nature of subscenes does not distinguish primary
scenes from other frame-like mental structures. It would perhaps be more
expedient to think of frame-like representations as being composed of either
individual entities and relations, combinations of entities and relations, com-
plex events consisting of simple ones (such as cooking or speaking), or
subdomains constituting larger domains. For the purposes of the study of
metonymy, we could conceive all of them as elements of frame-like structures
that can give rise to metonymies. The elements can be of any complexity and
either one can be used for another element or the whole (given the conditions,
including “sensory over nonsensory,” mentioned in the previous subsection).
However, it is clear that for metaphor to emerge from such frame-like struc-
tures we need two elements that have more complexity than simple entities, as
Grady and Johnson correctly observe. But this complexity may derive not only
from subscenes but also from any complex events or subdomains.
The Nature of Correlations

Another question we should consider in connection with the “metaphor-
metonymy relationship” issue is what we mean by correlation. Proponents of
the theory of primary metaphors appear to limit correlations to dimensions of
experience that are “tight” and “which unfold dynamically over very brief time
spans.” There are, however, many other correlations in experience that are less
tight and that occur over much longer time spans. Such less tight and less brief
correlations in experience can also bring together elements (of any complexity)
in frame-like mental structures and can, at least initially, produce metonymies
within the frame, and, then, metaphors that are based on the metonymies. It
might turn out that some of our concepts that form conceptual metaphors today
may be (or may have been) based on correlations in experience that are less
tight and that have worked through historical time (e.g., correlations between
symbols and their referents, between models and what they are models of ).
The evolution of particular cultures and systems of categories can probably
also produce metaphors that emerge over long time spans through a meto-
nymic stage in the process. If this suggestion is valid, a number of conceptual
metaphors that are now taken to be nonprimary and not based on correlations
may turn out to be correlational ones, thus further blurring the distinction
between primary and nonprimary metaphors.
3.4 Conclusions
In the chapter, I suggested that many metaphors arise from, and are not
independent of, metonymies. It should be made clear that, contrary to some
interpretations, the claim is not that correlation metaphors are not really
metaphors but metonymies. For example, Dancygier and Sweetser (2014:
104) write in connection with my suggestion: “Kövecses (2013) explicitly
argues that all conceptual metaphors are based in metonymy; however, correl-
ation is not metonymy but the basis for metonymic uses of one thing to stand
for something correlated with it.” This is not what I claim. First, I do not claim
that all conceptual metaphors are based in metonymy. Obviously, resemblance
metaphors are not metonymy-based. Second, my proposal was that correlation-
based metaphors emerge from frame-like mental representations through a
metonymic stage (not a metonymic linguistic expression), a “metonymy-pro-
ducing relationship” (see Kövecses and Radden 1998; Radden and Kövecses
1999) between two elements of a frame (if you wish, “the basis for metonymic
uses”). This is, I assume, a perfectly natural process. It is in line with how
correlations in general produce a large number of metonymy types. Indeed, we
can think of experience types as being constituted by several more basic
correlated elements (parts). Given the elements of experience constituting
experience types, there are a large number of metonymy-producing relations
that underlie the metonymies (metonymic expressions) we use.
When do the elements that are metonymically related lead to metaphors?
I proposed this happens when one of the elements of a frame-like mental
structure is generalized (schematized) to a concept that lies outside the initial
frame in a different part of the conceptual system. The generalization process
leads to sufficient conceptual distance between the initial and the new frame on
which metaphors can be based. In order for metaphors to emerge, in the typical
3.4 Conclusions 49
cases, distant vertical taxonomies are involved. This explains the importance
of taxonomic hierarchies and frames in our conceptualization of the “meta-
phor-metonymy relationship.”
The generalization (schematization) process can apply to either the element
in a frame that becomes the source or the target domain of a metaphor. When it
applies to an element that becomes the source (Case 1), the other element will
be the target of the metaphor (e.g., sadness is down). I described this
situation as “target giving rise to the source.” In the other case (Case 2), the
generalization applies to an element that becomes the target, and the other
element will be the source (e.g., knowing is seeing). I described the
situation as “source giving rise to the target.” Furthermore, when either the
source or the target is generalized (schematized), we can get elaborations. For
example, heat as the source of anger can be elaborated as “fire,” “stew,”
“boiling water,” “blood,” “piss,” “volcano,” “lava,” and so on. And when, say,
difficulty as target is elaborated in the metaphor difficulty is weight,
we get “problems,” “emotions,” “responsibilities,” and so on (and hence we
get the metaphors problems are weights, emotions are weights, and
responsibilities are weights). Needless to say, the source weight can
also be elaborated (as, for instance, “burden” and “baggage”) – in, often,
culture-specific ways.
I believe the view outlined here provides a coherent overall account of why
many metaphors can be interpreted as metonymies and why many metonymies
can be taken to be metaphors. The reason seems to be that, in the case of
correlational metaphors, between a nonfigurative and a metaphoric stage there
is a metonymic one. Given a frame in which the correlations apply between
two elements, there must be a logically prior metonymic relationship between
the elements before one element is moved out of the initial frame to establish a
new and independent frame, which will form either a metaphoric source or
target. Only future research can decide whether this explanation can be
extended to cases of conceptual metaphors not analyzed (only hinted at) in
the present study.
4 Domains, Schemas, Frames, or Spaces?
In the previous chapter, I have talked about “frame-like” conceptual structures

in connection with metaphor. This way of talking about the conceptual struc-
tures that are involved in conceptual metaphors is useful when we wish to
highlight what is common to such structures (namely, that they are all coherent
organizations of experience). However, if we are interested in the variation of
the structures that conceptual metaphors can utilize and display, we need to be
able to identify the way (or ways) in which these conceptual structures
systematically differ from one another.
Metaphor researchers, including myself (see Chapter 3), working within the
framework of cognitive linguistics use a number of different terms to refer to
the conceptual structures that they assume to constitute conceptual metaphors.
The most commonly employed term is that of domain (as in source and target
domain), but several others are also in circulation, including image schemas
(e.g., Lakoff 1990, 1993), frames (e.g., Kövecses 2006; Lakoff 1996), scenes
(e.g., Grady 1997a,b), mental spaces (e.g., Fauconnier and Turner 2002),
schemas (e.g., Lakoff and Turner 1989), and scenarios (e.g., Musolff 2006,
2016). There are more terms in use (such as model), but the list is sufficient to
indicate some terminological confusion among the practitioners of conceptual
metaphor theory. But more importantly, the terminological chaos is a reflection
of a serious, deep-seated theoretical-conceptual dilemma; namely, the diffi-
culty of identifying the appropriate conceptual unit, or structure, that partici-
pates in the formation of conceptual metaphors. The present chapter is an
attempt to offer a way of handling this issue.1
One could suggest of course that these designations and the conceptual
content corresponding to them are for all practical purposes equivalent, and
that they can replace one another in versions of CMT. Or one could argue, as
Grady (1997a) and Musolff (2006) do (for different reasons), that domains are
not the appropriate conceptual structures where metaphorical conceptualiza-
tion takes place. Would this mean that we can discard the notion of domain as
1
In the first half of the chapter, I reuse materials from Kövecses (2017b).
50
4.1 Image Schemas, Domains, Frames, and Spaces 51
a theoretical construct in CMT? Finally, one could select just a pair of these
terms (such as image schema and domain or mental space and domain) and
claim that these are sufficient to describe and explain conceptual metaphors. It
seems then that there is a very real need to raise the issue of which conceptual
structure or structures (or unit(s)) constitutes (or constitute) conceptual meta-
phors. Given this situation, my main goal in the chapter is to answer the
following general question:
What is (are) the appropriate conceptual structure/unit (or structures/units)
involved in conceptual metaphors?
My suggestion will be that it is best to think of conceptual metaphors as
simultaneously involving conceptual structures, or units, on several distinct
levels of schematicity (see, e.g., Lakoff 1987; Langacker 1987; Rosch 1978).
I distinguish four such levels: the level of image schemas, the level of domains,
the level of frames, and the level of mental spaces (in addition to the linguistic
level of the actual utterances in which the metaphors are instantiated). My
claim is that the resulting overall picture of conceptual metaphors provides us
with a new comprehensive framework for the study of metaphor in conceptual
metaphor theory. I would like to refer to the emerging new framework as the
“multilevel view of conceptual metaphor.” To demonstrate my points, I will
mainly use the source domain of building in this chapter (see, e.g., Grady
1997a,b; Kövecses 1995, 2000a,b, 2002/2010; Lakoff and Johnson 1980).
In what follows, first, I describe the theoretical background, where I justify
the relatedness of the four levels mentioned above, independently of metaphor.
Second, based on these findings, I apply the results to the study of conceptual
metaphors in general. Third, I look at the well-studied source domain of
building in some detail in order to show how the levels work together in
this particular example. Fourth, I summarize the main points and describe the
general picture of the multilevel view of metaphor. Fifth, I examine some of
the consequences of the new view for a variety of issues related to metaphor.
Finally, I offer some conclusions, together with indicating some potential
benefits and future directions of research involved in the new model.
4.1 Image Schemas, Domains, Frames, and Spaces

Following Lakoff (1987), Langacker (1987), and Rosch (1978), I assume that
much of our knowledge about the world comes in large systems of concepts
organized into hierarchies at various levels (superordinate level, basic level,
subordinate level). The systems of concepts so connected are related by
schematicity, which is defined by Langacker (1987: 492) as follows: “Sche-
maticity [is] Relative precision of specification along one or more parameters.”
If this is the case, we can see the “source domains” of conceptual metaphors as
a large system of concepts related to each other at several levels, or layers, of

precision of specification. The view entails that, given particular concepts, the
various levels of schematicity form a continuous hierarchy; the various levels
shade gradually into more or less schematic levels. In other words, the levels
within such schematicity hierarchies do not have rigid boundaries but are
graded as regards their schematicity.
I propose that the conceptual structures mentioned in the introduction (i.e.,
image schema, domain, frame) can be regarded as occupying different levels in
such schematicity hierarchies. Specifically, given such hierarchies, we can
distinguish fairly clearly four different levels of schematicity, each designated
by the four terms: image schema, domain, frame, and mental space. The four
levels go from the most schematic to the least schematic, as represented in the
diagram below:
Most schematic: image schema
domain
frame
Least schematic: mental space
Figure 4.1 Schematicity hierarchy for four conceptual structures

(Dancygier and Sweetser 2014, propose a similar hierarchy)
The upward arrow in the diagram indicates increasing schematicity, whereas
the downward arrow indicates increasing specificity. The schematicity-based
hierarchy can also be represented as a series of inclusions, as in the
diagram below:
IS
DM
FR
MS
Figure 4.2 Schematicity as inclusion

In the diagram, IS stands for image schema, DM for domain matrix, FR for
frame, and MS for mental space.
Image schemas are essential conceptual structures that imbue experience
with meaning (cf. Johnson 1987; Lakoff 1987). Based on the literature on
image schemas, Hampe (2005: 1–2) finds four features of image schemas
especially characteristic. I present them here in a somewhat simplified form.
Image schemas are
directly meaningful preconceptual structures;
highly schematic gestalts;
continuous analogue patterns;
internally structured, consisting of only a few parts.
Because of their highly schematic nature, image schemas range over the entire
conceptual system making a wide variety of concepts and experiences mean-
ingful. For example, the concept of journey presupposes the more schematic
structure of motion and, more specifically, source-path-goal motion
(to distinguish it from other types of motion). Further, concepts may take
several image schemas to support them conceptually. For example, the concept
of body is based on the image schemas of container, verticality,
(structured) object, and so on. Finally, concepts may also be character-
ized by the same image schemas. For instance, the concept of building (in
the sense of an enclosed construction), similar to the body, presupposes the
container, verticality, and object schemas.
The notion of domain is defined by Langacker (1987: 488) as follows:
“Domain [is] A coherent area of conceptualization relative to which semantic
units may be characterized.” Unlike image schemas, domains are not analogue,
imagistic patterns of experience but propositional in nature in a highly sche-
matic fashion. They are at a level immediately below image schemas. Image
schemas make domains such as journey, body, and building (mentioned
above as concepts) meaningful. The various image schemas apply to different
aspects or dimensions of a domain in the sense of a domain matrix (Langacker
1987). A domain as a domain matrix (such as building) presupposes a
variety of concepts that characterize different aspects of the domain. Domains
have many more parts than image schemas, and are thus more information-
rich. The definition of domains does not distinguish between domains and
frames, or idealized cognitive models (Lakoff 1987), as noted by Langacker
himself. The only way to distinguish the two is in terms of schematicity.
Frames (see, e.g., Fillmore 1982) are less schematic conceptual structures
than domains. As noted above, the definition of domain by Langacker can
serve as a definition of frame as well. (On the blurred boundaries between
domains and frames, see also Cienki 2007.) The difference between a domain
and a frame, in my view, can be captured by a difference in schematicity
between the two: Frames elaborate particular aspects of a domain matrix; that
is, particular higher-level concepts within a domain. Sullivan (2013) considers
this relationship as inclusion. According to her, domains include or consist of
frames. We can also think of the relationship between the two as different
degrees of schematicity. It can be suggested that frames involve more concep-
tually specific information than domains. For example, the body domain can
be seen as being elaborated by several distinct frames, such as perception,
ingestion, and exercising (see Sullivan 2013). These frames account for
such metaphorical linguistic expressions as I see what you mean (percep-
tion), digest an idea (ingestion), and a mental exercise (exercising)
(Sullivan 2013). Together, they make up what is known as the generic-level
metaphor the mind is the body (see Johnson 1987; Sweetser 1990). In
general, the frames elaborating a domain consist of roles and relations between
the roles and the roles can be filled by particular values.
When the roles are filled by particular values in actual discourse in specific
communicative situations, we have to do with mental spaces (see Fauconnier
1994). I use mental spaces in the sense the term is defined by Fauconnier
(2007: 351): “Mental spaces are very partial assemblies constructed as we
think and talk, for purposes of local understanding and action. They contain
elements and are structured by frames and cognitive models. Mental spaces are
connected to long-term schematic knowledge, such as the frame for walking
along a path, and to long-term specific knowledge, . . .” This is also what
Langacker (2008) calls “current discourse space” (CDS). In Langacker’s
words: “The CDS is a mental space comprising everything presumed to be
shared by the speaker and hearer as the basis for discourse at a given moment”
(Langacker 2008: 59). Mental spaces borrow their structure from frames, but
the generic structures from frames are further elaborated by specific infor-
mation from context. In addition, mental spaces can be structured by one or
several different frames. That is, they can be the realizations of a single frame
or they can rely on a combination of roles and relations from several distinct
frames. Mental spaces are, then, even more specific than frames, in that they do
not operate with generic roles and relations in most cases but with specific
instances of roles and relations. At the same time, they are also coherent
organizations of experience, just like frames and domains, but they function
at a very specific and conceptually rich level. Moreover, mental spaces are
used in online processing for purposes of local understanding, as noted above,
and also by Fauconnier and Turner (2002), but, as we just saw, they are
structured by frames. Mental spaces are online representations of our under-
standing of experience in working memory, whereas frames and domains are
conventionalized knowledge structures in long-term memory.
In brief, image schemas, domains, frames, and mental spaces are all used
by conceptualizers/speakers for the purposes of lending organization and
coherence to our experience. I take image schemas to be the most schematic

and mental spaces the least schematic (i.e., the conceptually richest) cogni-
tive structures that can have this organizing function. As we will see in the
next section, they all can, and do, play a role in metaphorical conceptual-
ization. (In addition to these four levels, there is of course the level of
communication, Level 5, where speaker and hearer use some symbols (lin-
guistic or otherwise) that make manifest, or elaborate, the content of particu-
lar mental spaces.) We can represent their relationship in a schematicity
hierarchy (SH) as follows:
SH 1 SH 2 SH 3 SH 4 Schematicity hierarchies
L1 Image schemas
L2 Domain matrixes
L3 Frames
L4 Mental spaces
L5 Utterances used in actual communication
Figure 4.3 The schematicity hierarchy with three major distinctions between
the levels
The diagram represents the four levels of conceptual structures in four

different schematic hierarchies (SH), one of which could be, for instance, the
SH for building. The dotted line that separates image schemas and
domains/frames is intended to indicate that image schemas are analogue
structures, whereas domains and frames are not. The bold dotted line between
domains/frames and mental spaces indicates that domains and frames
are in long-term memory, whereas mental spaces are used in online processing
in working memory, as noted by Fauconnier above. The dotted line between

L4 and L5 separates the conceptual from the linguistic.
4.2 Metaphor and Structured Conceptual Experience

We can now relativize this theoretical background to the treatment of meta-
phor. In order to get clear about the general issue of how schematicity plays a
role in metaphorical conceptualization, below I examine the relationship
between image schemas, domains, frames, and mental spaces in more detail
in relation to metaphorical conceptualization.
Let us begin with image schemas. Conceptual metaphor theory has been
instrumental in resolving one of the major issues in the theory of cognition:
How can we account for the meaning of abstract concepts? After all, abstrac-
tions are what they are since we do not have direct physical, perceptual
experience related to them. The proposal that CMT made is that abstractions
(abstract concepts and reasoning) are made meaningful by virtue of conceptual
metaphors connecting an abstract concept with a physical (i.e., perception-
based) one (see, e.g., Barsalou 1999). In them, extremely general patterns of
perceptual experience called image schemas give meaning to abstract concepts
(see, e.g., Johnson 1987; Lakoff 1987, 1990). As a result, the conceptual
system, including abstract concepts, becomes embodied, that is, perception-
based through and through.
Conceptual metaphors in CMT are usually described as relationships
between domains. However, as it has often been noticed, to think of conceptual
metaphors as a set of mapping relations between two domains leads to a major
problem: Source domains typically contain a lot more conceptual material than
what is actually mapped onto the target domain. Several solutions to the
problem have been suggested (e.g., Clausner and Croft 1997; Grady 1997a,
b; Kövecses 2000a; Lakoff 1990) and, with the exception of Grady (who
postulates primary scenes, as we saw in Chapter 3), they all rely on the notion
of domain. The solutions either try to constrain the transfer of source domain
materials by means of the schematic structure of the target (e.g., Lakoff 1990)
or try to narrow the source domain itself that participates in the mapping
(Kövecses 2000a, 2002/2010). Why did the problem of overestimation arise
(in choosing domains)? In all probability, it did because researchers had a very
broad notion of domain in mind – a notion defined, as we saw above, by
Langacker (1987: 488) as “A coherent area of conceptualization relative to
which semantic units may be characterized.” If we think of all the concepts that
are simultaneously presupposed by a given concept, that is, a domain matrix, it
is clear that domains (in the sense of domain matrix) are highly inclusive
knowledge structures. Not all concepts in, that is, not all aspects of, a domain
matrix, are utilized in metaphorical conceptualization. The domain matrix for,
4.2 Metaphor and Structured Conceptual Experience 57
say, building, as we will see in the next section, represents our entire
knowledge structure associated with this concept. In general, out of the entire
domain matrix, only select aspects of the domain matrix participate in the
mappings between source and target concepts.
The aspects that do participate in the mappings can be, and usually are,
given in the form of frames – at a lower level of schematicity. The frames
elaborate the select aspects of domains. Thus, the source frames offer more
specific information than domains, but they do not cover or exhaust all aspects
of a source domain (matrix).
As regards metaphorical mental spaces, they are the least schematic con-
ceptual structures of the four discussed here (i.e., image schemas, domains,
frames, and mental spaces). They are highly specific structures occurring in
online processing in particular communicative situations. As such, they are
fully influenced by several different kinds of contextual information (Kövecses
2015a). Mental spaces participating in metaphor can be seen as elaborations of
frames that elaborate select aspects of domains for metaphorical conceptual-
ization. They contain the most specific information that derives from filling out
generic roles with particular values, as well as from the specific context in
which the spaces emerge.
While domains and frames belong to the level of cognitive organization and
analysis of metaphor that Kövecses (2002/2010) calls the “supraindividual”
level, mental spaces occupy a level that can be thought of as the “individual”
level. In a previous publication (Kövecses 2002/2010), I distinguish the
supraindividual, individual, and subindividual levels in the following way:
“In conclusion, then, the cognitive linguistic view of metaphor that has been
discussed in this book works on three levels: the supraindividual level corres-
ponding to how a given language and culture reflects decontextualized meta-
phorical patterns, the individual level corresponding to the metaphorical
cognitive system as used by individual speakers of a language, and the
subindividual level corresponding to universal aspects of various kinds of
embodiment.” (Kövecses 2010: 321)
The supraindividual level corresponds to the level of domains and frames. In
the CMT literature, descriptions of conceptual metaphors are typically given at
the supraindividual level. Such descriptions lack the specificity and richness of
metaphorical conceptualization that we find at the individual level – the level
corresponding to the use of mental spaces in the present framework.
The idea that mental spaces can be thought of as the most specific concep-
tual structures participating in metaphorical conceptualization finds some
support in more recent work. Musolff (2006) uses the term “scenario” in the
same sense in which I use “mental space.” (Actually, Musolff also roughly
equates his use of scenario and scene with mental space.) In discussing a
particular example, he writes:
The source schema of these two examples, MARRIAGE PARTNERS = PARENTS OF

CHILD, may be simple, but the characterization of the participants in terms of their
roles, intentions, and states of minds, as well as the assessment of their actions in terms
of chances of success, are in fact highly specified. The readers are not only provided
with a general schematic frame to understand the order of events and a few causal links
between them, but rather with a whole little scene, complete with the presumed
“interests” and “biases” on the part of the participants and an evaluative interpretation.
(Musolff 2006: 27)
In other words, Musolff also suggests that scenes and scenarios function below
the level of domains and frames, where speaker and hearer metaphorically
conceptualize their experiences in a fully contextualized fashion. The speaker
and hearer take advantage of a large amount of knowledge available to them in
the specific communicative situation. His observations are thus very much in
line with the present approach that emphasizes the various degrees of sche-
maticity in metaphorical conceptualization.
We can represent the emerging picture of metaphorical conceptualization in
a schematic way in the diagram to follow:
Source Target
Figure 4.4 Same-level mapping
The diagram indicates that image schemas metaphorically structure the

most schematic concepts and mental spaces (or scenarios) the least
schematic ones.
4.3 The Example of building 59
4.3 The Example of building

Let us take the source domain of building as an example to demonstrate
some of the points above and highlight some additional ones. The building
source domain has been studied by a variety of authors (see, especially, Grady
1997a,b; Kövecses 1995a,b, 2000a, 2002/2010; Lakoff and Johnson 1980)
from several different perspectives. In what follows, I will mostly rely on my
own work.
building is an extensive domain with a large number of members (house,
cathedral, garage, etc.). Instead of giving the concept of building a
dictionary-like definition (such as, for instance: Building is a man-made
structure typically with walls and a roof that enclose an internal space and
that separate and protect it from an external space), in line with the argument in
this chapter, we can describe it at various levels of schematicity in the
following way:
As noted above in section 2, building is a concept that is based on, or
characterized by, several image schemas, such as container, vertical-
ity, part-whole, and object.
The building domain matrix is constituted by a large number of schematic
concepts. Specifically, the building is a physical object in space; as a
physical object it has size; it has a location; it has both vertical and
horizontal extension; it is a structured object; it is a man-made
structured object; it is made of physical matter; it is an enclosed space
that is partitioned into parts; it has a particular shape and color; it has
some function; and so on.
In addition, the building domain consists of a number of frames. Most
obviously, it possesses a building (as a process) frame; a building is
something that needs to be built. It also possesses a (physical) support
frame, in which there are certain elements organized in a structured way.
Furthermore, it has a frame for the parts of the building, such as walls,
rooms, doors, windows, chimneys, basement, cellar, attic, roof, floors, stairs,
and many others. Moreover, the domain of building consists of a function
of building frame that provides information about who use the building, in
what ways, and for what purpose(s). There are no doubt several other frames
that make up the domain of building, but for the present illustrative purposes
these four will do: building, (physical) support, parts of building,
and function of building.
At the mental spaces, or scenario, level, we would have further specifica-
tions of any of the frames discussed above. For example, an elaboration of the
building frame would be the mental space associated with a sentence like:
“John built himself a strong house.” The sentence would specify the specific
individual who built the house and a structural property of the house. These
are pieces of information that are not contained in the more schematic
building frame.
Now let us look at the commonly cited metaphorical examples that can be
found in the CMT literature in connection with the building source domain.
I will mainly use here what I call the “lexical approach” to collecting metaphors
that are characteristic of a particular domain such as building (see Kövecses
1986, 2015b). In essence, the lexical approach takes advantage of a variety of
lexical materials (idioms, collocations, synonyms, etc., as these are available in
dictionaries and other collections of lexical items) in finding metaphorical
expressions. The method performs well as regards collecting conventionalized
and decontextualized types of metaphorical expressions; it performs poorly as
regards finding novel types and tokens of metaphors and quantifying the
metaphorical data in relation to a domain. Since in the chapter I am concerned
with the conventional types of metaphorical expressions (and not with quanti-
fying metaphor occurrence in actual use), I employ the lexical method (see also
the subsection on methodology discussed later in this chapter). Nevertheless,
I will also rely on some corpus examples in the analysis to follow.
We can begin with Lakoff and Johnson’s Metaphors We Live By, where we
find this for the metaphor theories are buildings (Lakoff and Johnson
1980: 46):
theories (and arguments) are buildings
Is that the foundation for your theory?
The theory needs more support.
The argument is shaky.
We need some more facts or the argument will fall apart.
We need to construct a strong argument for that.
I haven’t figured out yet what the form of the argument will be.
Here are some more facts to shore up the theory.
We need to buttress the theory with solid arguments.
The theory will stand or fall on the strength of that argument.
The argument collapsed.
They exploded his latest theory.
We will show that theory to be without foundation.
So far we have put together only the framework of the theory.
For the related concept of argument, we find the following examples:
an argument is a building
We’ve got the framework for a solid argument.
If you don’t support your argument with solid facts, the whole thing
will collapse.
He is trying to buttress his argument with a lot of irrelevant facts, but

it is still so shaky that it will easily fall apart under criticism.
With the groundwork you’ve got, you can construct a pretty strong
argument.
In the Collins Cobuild’s English Guides by Deignan (1995), the following
examples (i.e., linguistic metaphors) can be found for theories are build-
ings (taken from Kövecses 2002/2010):
theories are buildings
Increasingly, scientific knowledge is constructed by small numbers of
specialized workers.
McCarthy demolishes the romantic myth of the Wild West.
She lay back for a few moments contemplating the ruins of her
idealism and her innocence.
Don’t be tempted to skip the first sections of your programme, because
they are the foundations on which the second half will be built.
. . . the advance that laid the foundations for modern science.
Our view, he said, is that these claims are entirely without foundation.
My faith was rocked to its foundations.
The second half of the chapter builds on previous discussion of
change and differentiation in home ownership.
Given such examples as the above, I suggested that the theories are
buildings conceptual metaphor focuses on, or profiles, three aspects of the
concept of theory (Kövecses 2002/2010): (1) the aspect of the construction of
a theory (as exemplified by build, construct, put together); (2) abstract structure
(as exemplified by (without) foundation, groundwork, framework, build on, lay
the foundations, collapse, demolish, shore up, buttress); and (3) abstract stability
or lastingness (as exemplified by strong, solid, shaky, shore up, buttress, fall
apart, in ruins, rock to its foundations, stand or fall). (As can be seen, some of the
linguistic examples can characterize several different aspects of a concept simul-
taneously.) In addition, I proposed that such “meaning foci” (that is, the habit-
ually profiled aspects) of building as creation, abstract structure, and
lastingness can be captured by means of mappings. Finally, I also suggested
that the building source domain has a wide scope (see, Kövecses 2000a,b)
and applies to “complex abstract systems” in general, not just to the concept of
theory. Here are the three mappings corresponding to the three meaning foci:
building ➔ creation or construction of a complex abstract system
physical structure of the building ➔ abstract structure of a complex
abstract system
physical strength (of the structure to stand) ➔ abstract stability/
lastingness (of the complex abstract system)
Essentially, it is these meaning foci that, with the exception of the first,
Grady’s (1997a,b) primary metaphors also capture: logical organiza-
tion is physical structure (corresponding to 2) and persistence is
remaining erect (corresponding to 3). In other words, if this analysis is
correct, we can suggest that primary metaphors and the main meaning focus
are, in part, related phenomena: When the mappings of a conceptual metaphor
are based on primary metaphors such as the ones above, these will determine
the main meaning focus of the complex metaphor (like theories are
buildings). (However, it should be noted that the notion of main meaning
focus also applies to non-primary-metaphor-based, i.e., analogy-based, con-
ceptual metaphors, such as computers are people or people are com-
puters, where people in the former metaphor are associated with, e.g.,
emotionality, while computers in the latter with, e.g., a high level of memory
capacity and mechanical calculations.)
In light of the analysis so far, it seems that the building source domain
participates in the conceptualization of theories (and, more generally, of
complex abstract systems) by means of two frames within it: the
building (as process) frame associated with the construction aspect
and the (physical) support frame with the structure aspect of the
building domain. The building (as process) frame gives rise to the first
mapping above, “building ➔ creation.” The frame utilizes the conceptual
elements of the builder, the action or activity of building, and the thing that
gets built. Thus, the builder corresponds to the person who creates or
constructs the theory, the building action to the creation or construction
process that results in the theory, and the thing built to the theory itself.
Interestingly, it does not conventionally utilize the materials that are used
in the building activity – at least according to the examples given (but
see below).
The second frame, (physical) support, is represented by the second
and third mappings above. In these mappings, we have two important
structural elements that are related to each other in a particular way, as well
as a property that is expected to characterize both the elements and the
structural relationship between them. The structural elements are the founda-
tion and the outer shell (of the building), and the relationship that obtains
between them is that the outer shell is supported by the foundation. The
property that characterizes the elements and the relationship between the two is
that of strength. These are the conceptual elements that are utilized in the
(physical) support frame.
The emphasis on this latter aspect of buildings has a long cultural history.
It is this relationship (supported by) and this property (strength) that are
emphasized, for example, by one of the parables in the New Testament:
(1) 25 The rain came down, the streams rose, and the winds blew and beat against that
house; yet it did not fall, because it had its foundation on the rock. 26 But everyone who
hears these words of mine and does not put them into practice is like a foolish man who
built his house on sand. 27 The rain came down, the streams rose, and the winds blew
and beat against that house, and it fell with a great crash.” (Matthew 7:25–27 New
International Version (NIV))
This may be a culturally significant factor that may have created, or, simply,
may reinforce, the conceptual use of these two specific frames within the
building source domain. The other two frames mentioned above, parts
of the building and function of the building, do not appear to
produce conventionalized metaphorical language on a par with building (as
process) and (physical) support, as evidenced by the examples used in the
literature and the dictionaries consulted.
The meanings of the conventionalized expressions, the three mappings on
which they are based, the two frames associated with the building source
domain, as well as the other constitutive frames are at, what I called, the
supraindividual level (i.e., in long-term semantic memory), that is, at the level
where we store decontextualized conceptual information related to building
as a conceptual domain.
The issue we consider next is what happens when this body of information
is actually used in real discourse at the individual level, that is, where actual
people communicate in real contexts characterized by a great deal of specific
information available to them.
The building Source Concept at the Individual Level

We have seen above that when we study building metaphors at the suprain-
dividual level (i.e., the level of domains and frames), we find that the frame
elements and the metaphorical expressions that are associated with them
emerge from two of the constitutive frames in the building domain: build-
ing (as process) and (physical) support.
When we make use of these frames in the course of actual cases of online
metaphorical conceptualization (i.e., at the individual level), speakers can
extend the use of frame elements to new ones. I mentioned above in connec-
tion with the theories are buildings (and more generally, complex
abstract systems are buildings) metaphor that the conventionally used
frame elements in the building (as process) frame involve the builder, the
building process, and the building that gets built. But, clearly, the building
materials are also elements within this frame. Given the data above, the
building materials are not conventionally used for metaphorical purposes. In
addition, since most rational activities (such as building an object) require
preparation and planning, the application of the building frame can also be
extended to such aspects of the building process. I assume that this is what
Lakoff and Johnson (1980/2003) refer to as the “extension of the used part” of
the metaphor. They exemplify it with the sentence “These facts are the bricks
and mortar of my theory,” where bricks and mortar constitute the frame
element of building materials. Additional examples are not difficult to find.
A simple Google search returned the following example: “This middle level
between theories and concepts allows models to serve a critical function within
science; they act as the bricks and mortar of a theory and are the basis for how
scientists argue” (www.ncbi.nlm.nih.gov/pmc/articles/PMC3671648/). The
use of the building frame with the extension to building materials may not
be frequent and conventional, but when speakers conceptualize theories meta-
phorically in actual discourse they may be prompted by the frame to extend its
application to this frame element (by way of completing the frame).
Lakoff and Johnson (1980/2003) also mention another example: “His theory
has thousands of little rooms and long, winding corridors.” They refer to it as a
case in which the metaphor derives from an “unused part” of the source
domain. We can now provide a fuller account of such cases within the
framework I am developing here. It can be suggested that in cases like this
frame elements from other than the routinely or conventionally used frames
can participate in metaphorical mappings. The cognitive mechanism that
seems to be involved can be characterized as follows. Again, Musolff (2001)
provides several pertinent examples, which I reanalyze here (based on
Kövecses 2005):
“We are delighted that Germany’s unification takes place under the European roof.”
(Documentation by the Federal press- and information office, Bonn)
“At the moment, the German occupants of the first floor apartment in the ‘European
house’ seem to think that foreigners from outside the continent should be content with
living in the rubbish bin.”
“What does he [Chancellor Kohl] need this house for, after so many years as
Chancellor? – Well, it’s obvious, he wants to become the caretaker.” (Die Zeit, May
16, 1997)
“[the European house is] a building without fire-escapes: no escape if it goes wrong”
(The Guardian, May 2, 1998)
“[it is a] burning building with no exits” (The Times, May 20, 1998)
Musolff notes that, given the common european house metaphor intro-
duced in public discourse in the 1990s, i.e., the european union is a
building metaphor, all kinds of metaphorical expressions showed up in
political discourse at the time that were different from the ones presented at
the beginning of the section (i.e., were outside the profile, or meaning focus, of
the source). In addition to the highly conventionalized linguistic metaphors,
such as lay the foundations, building, buttress, framework, and so on,
journalists and others used novel, unconventional metaphors, such as roof,

apartment, caretaker, and fire-escape. In my own analysis of these new,
unconventional examples (Kövecses 2005), I pointed out that given the main
meaning foci of the building source domain, these unconventional meta-
phorical expressions should not have occurred. Nevertheless, I provided the
description of a cognitive mechanism that might account for the comprehen-
sion of these unexpected items in real discourse (see Kövecses 2005). I argued,
briefly, that once a source domain (e.g., building) is used to structure a target
(e.g., in the present case, the european union) by means of conventional
source elements (i.e., those in its foci), elements in the target so conceptualized
may “select” elements of the source that fit that target element (but are not
conventionally associated with it). Although there is no conventional corres-
pondence, say, between the possibility of leaving the EU in the target and the
fire-escape in the source, the target element may “select” the source element if
it “fits” it. (The cognitive operation that produces such examples can also be
regarded as a case of conceptual integration. See Fauconnier and Turner 2002
and Chapter 6.)
In light of the suggestions in this chapter, I would like to add a missing piece
of the argument to that offered in 2005. As was noted above, at the supraindi-
vidual level the source concept of building utilizes the frames of building
(as process) and (physical) support for establishing mappings between the
building domain and the target domain of complex abstract systems
(including theories and political systems, such as the EU). At the
individual level, however, it appears to additionally involve both the parts
of the building and the function of building frames. Roof, apart-
ment, and fire-escape are component parts of buildings, while caretaker has to
do with the function of buildings (namely, the function of providing living or
working space for people, and that, given this function, a caretaker handles the
daily maintenance and operation of the building on behalf of the people living
or working there). Thus, while at the supraindividual level of metaphorical
conceptualization certain specific frames within a source domain are profiled
(and they are routinely or habitually used for a particular purpose), at the
individual level we find additional frames that may be involved in performing
the task. The supraindividual level seems to constrain the use of frames within
a domain, whereas the individual level seems to broaden their use. The
difference no doubt results from the lack, as opposed to the presence, of
information-rich context (as described in Kövecses 2015a) and the more
flexible use of frame elements (through frame completion).
To summarize the main points, the following diagram indicates how
building as a source concept participates in metaphorical mappings in
relation to the complex abstract systems target (IS = image schema;
DM = domain; FR = frame; MS = mental space):
BUILDING COMPLEX ABSTRACT SYSTEMS

(CAS)
IS
CAS
EXTENSION CONTAINER OBJECT WHOLE-PART
DM SHAPE SIZE COLOR CREATION ABSTRACT

ABSTRACT CREATION
VERTICAL SPACE LOCATION PARTS
PARTS ABSTRACT
HORIZONTAL PHYSICAL ENCLOSED FUNCTION FUNCTION
MATTER ABSTRACT
PHYSICAL STRUCTURE ABSTRACT
OBJECT SYSTEM
STRUCTURE
FR PHYSICAL
BUILDING SUPPORT CONSTRUCTION LOGICAL SUPPORT
architect foundation designer framework/organization
PARTS OF plans FUNCTION OF supports designs is based on/supported by
BUILDING builder BUILDING shell builder basic ideas
builds strong builds supported in a stable way
building organization/framework
? from materials ? from ideas
MS MARY BUILDING JOHN’S HOUSE MARY BUILDING JOHN’S LIFE

A HOUSE WITHOUT FOUNDATION HER CAREER WITHOUT FOUNDATION
Figure 4.5 The complex abstract systems are buildings metaphor

The question marks below the parts of building and function of
building frames are intended to indicate that, except for some special circum-
stances, there are no conventionally (routinely) applicable mappings between
these frames and the corresponding frames in the target (parts of a specific
complex abstract system and function of a specific complex
abstract system, which latter frames are not shown in the diagram).
Some further explanation is required for the examples of mental spaces in
the diagram. At the mental spaces level, a large number of mental spaces could
be created that involve the various elements of the building (as process)
frame or the (physical) support frame. For this reason, we can only offer
here some potential examples of mental spaces for which a building-related
frame provides the structure. First, consider two non-metaphorical statements
related to building: “Mary is building her house” and “John’s house has no
foundation.” Now take the corresponding metaphorical statements “Mary is
building her career” and “John’s life is without a foundation,” as used, say, in a
conversation. The first metaphorical sentence creates a mental space that is
structured by the building (as process) frame. We can call this the mary
building a career space. The second metaphorical sentence creates a
different mental space: john’s life without foundation. This is based
on the (physical) support frame. The two examples illustrate how mental
spaces can elaborate particular frames.
4.4 The Multilevel View: The General Picture 67
The diagram above shows both the schematicity hierarchy of the source and
target concepts (from IS to MS) in a vertical arrangement and the mappings
between the source and target concepts on the four levels in a horizontal
arrangement. The concepts constituting the conceptual metaphors on the
various levels range from most schematic to least schematic. Thus we get the
following conceptual metaphors:
IS level:
complex abstract systems are objects
complex abstract systems are vertical objects
complex abstract systems are container objects
complex abstract systems are whole objects with parts
DM level:
a (complex abstract) system is a physical object (building)
the creation of the system is the physical creation of the
building
the structure of the system is the physical structure of the
building
the parts of the system are parts of the physical object (building)
FR level:
the construction of the system is building (as process)
logical support is physical support
submetaphor:
the stability of the building is the strength of physical support
MS level:
(in the form of particular examples:)
mary building a career is mary building a house
john’s life without a foundation is john’s house without a
foundation
These mappings at the different levels license the use of metaphorical expres-
sions at different levels of schematicity. For example, the word inside, as in
“inside the system,” relates to the image schematic level (the container
image schema), while the word structure (when used metaphorically) is based
on a mapping at the domain level. Similarly, the word builder in the phrase
“the chief builder of the organization” is related to the building (as process)
frame. In other words, metaphorically used words, just like their corresponding
conceptual metaphors, can be found at various levels of schematicity.
4.4 The Multilevel View: The General Picture

In light of the observations above, it seems reasonable to propose the following
view of how the conceptual system may be organized and how this system
participates in metaphorical conceptualization:
We can distinguish four levels of schematicity (vs. specificity). At the

highest level of schematicity, we have image schemas, such as container,
verticality, force, object, whole-part, and many others. These are
extremely schematic structures that arise from our most basic embodied experi-
ences. We use them when we encounter objects and events; they are our first
guides in conceptualizing experience. Image schemas function as source con-
cepts in extremely general conceptual metaphors like more is up, states are
containers, emotions are forces, events are objects, and com-
plex abstract systems are complex physical objects.
Below this level of schematicity, there are domains. The various domains are
conceptually supported by image schemas in the sense that the image schemas
apply to the various aspects of domains that characterize them. These aspects are
essentially the concepts that belong to the domain matrix associated with
building. For instance, the domain of building has the aspects of spatial
object, man-made structure, horizontal and vertical exten-
sion, construction, and so on. A domain represents a highly schematic
experience with a set of aspects organized into a meaningful whole. Domains
serve as source domains in such conceptual metaphors as the mind is the
body, complex abstract systems are buildings, emotions are
natural forces, time is motion, and so on. These conceptual metaphors
are characterized by more specific experiential content than the ones solely
employing image schemas. In this sense, domains elaborate image schemas.
At the next level “down,” we have frames. Frames further elaborate the
various aspects of domains. For example, the building (as process) frame
elaborates the construction aspect of the building domain and the
(physical) support frame elaborates the structure aspect of the build-
ing domain. The various aspects of domains consist of several distinct
elements, called “frame elements” in the terminology of frame semantics. In
other words, frames consist of different sets of frame elements, that is, sets of
entities and relations. Conceptual metaphors that are clearly based on frames
include love is fire, knowing is seeing, understanding is
grasping, anger is a hot fluid in a container, accepting is
swallowing, morality is physical strength, surprising some-
one is unexpectedly impacting someone, and many others.
I suggest that image schemas belong to what, in previous work, I called the
“subindividual level” and domains and frames are all conceptual structures that
can be found at the “supraindividual level.” Together, these constitute the
realm that informs the basic ontology of conceptual systems. These are the
structures that we have in long-term memory and that provide the conceptual
substrate of meaning in general and meaning in language in particular. All the
image schemas, domains, and frames put together comprise the conceptual
content of language and what we commonly refer to as encyclopedic
4.4 The Multilevel View: The General Picture 69
knowledge. Importantly, this is the kind of knowledge that can be retrieved

from decontextualized language (with the help of what I referred to as the
lexical approach to the study of metaphor above; see, Kövecses, 2015b), and,
hence, these conceptual structures are largely decontextualized themselves.
What’s Going on at the Level of Mental Spaces?

By contrast, when we use language in real communicative situations, we put this
huge amount of tacit knowledge to use in order to achieve particular goals
(social, expressive, rhetorical, etc.). The job is performed online by individual
speakers in specific contexts who manipulate and modify the conceptual struc-
tures in long-term memory according to their communicative needs and goals.
The conceptualization process and the language that is used are, in this case, fully
contextualized. It is at this level that we utilize mental spaces, or, equivalently,
scenes or scenarios, as suggested by Musolff, that are not part of our routinely
used cognitive and linguistic repertoire. They are fully specific (their elements
are values, not roles), they foreground elements not foregrounded routinely, they
bring in new elements from the same frame (or other higher-level structures),
they make new inferences and evaluations possible in context, they blend source
and target frames, and so on. We have seen a number of metaphorical examples
for several of these cases above. In sum, this is a large-scale mobilization of our
decontextualized, static, and routine conceptual structures in real-world meta-
phorical conceptualization in information-rich contexts by means of a variety of
cognitive processes that individual speakers perform online.
At the same time, my general proposal is that whenever we use metaphors
online at the mental spaces, or individual, level, we can only do this by relying
on the level of frames, domains, or image schemas, that is, the supraindividual
level. This means that the use of a particular linguistic metaphor in context is
linked to the entire system of connections (to mental spaces, frames, domains,
image schemas) in our knowledge about the world. I hypothesize that a
metaphor that is used in a specific communicative situation as part of a mental
space, or scene, will activate the frame structure to which it is linked, which
will, in turn, activate the domain of which the frame is a part, and the activation
will reach the image schema that conceptually supports the frame.
In general, this view is consonant with, and is in the same spirit as, a number
of others in the cognitive linguistic study of metaphor, such as Lakoff’s (1990)
“invariance principle” and Ruiz de Mendoza’s (1998) “extended invariance
principle.” Ruiz de Mendoza suggests that any kind of high-level structure is to
be preserved by lower-level structures. A further constraint on metaphor
production (the “correlation principle”) by Ruiz de Mendoza and Galera
(2014) states that a source domain should be selected in accordance with the
implicational structure of the target. In addition, the source domains are not
only coherent with their targets (or the other way around), but also with context
in a broad sense. Elsewhere, I proposed that many metaphorical source
domains are coherent with one or more contextual factors. This constraint
was called the “pressure of coherence” with the context (see Kövecses 2005,
2015a). There seems to be abundant evidence for this kind of activation in the
experiments conducted by Boroditsky, Casasanto, Gibbs, Colston, and other
researchers (see, e.g., Boroditsky 2001; Casasanto 2009; Gibbs 2006; Gibbs
and Colston 2012.) In their study of figurative language, Dancygier and
Sweetser (2014) make a proposal that is close to the present one as regards
the emphasis they place on different levels of schematicity in accounting for
metaphoric language. David, Lakoff, and Stickle’s (2016) notion of “cascades”
is a formalized way to capture the same phenomena. What distinguishes the
present approach from these is that (1) in addition to image schema, domain,
and frame, I include the notion of mental spaces in the hierarchy and (2)
I consider context to play an essential role in the emergence of the hierarchies.
(I discuss the role of context in Chapter 5.)
All in all, then, we have the following picture:
analogue
structure
IMAGE SCHEMAS
…………………………………………………………………………
long-term memory
DOMAINS
non-analogue structure
FRAMES
------------------------------------------------------------------------------------
MENTAL SPACES
working memory
Figure 4.6 Activation and structuring between the levels

The upper horizontal line indicates the division between analogue and non-
analogue conceptual structures. The lower, bold horizontal line indicates the
division between conceptual structures in long-term memory and those used
online in working memory. The vertical arrows indicate that higher schematic
structures can structure mental spaces and that the mental spaces can activate
higher-level schematic structures.
4.5 Some Consequences of the Multilevel View

In the present section, I wish to examine four consequences of the new,
multilevel view of conceptual metaphors in relation to (1) a proposal by
4.5 Some Consequences of the Multilevel View 71
George Lakoff, (2) a distinction emphasized by Gerard Steen, (3) the general
issue of metaphorical idioms, and (4) a fundamental issue in the methodology
of studying metaphor within a CMT framework. In particular, they are as
follows:
What are “deep” versus “shallow” or “superficial metaphors? What is their
ontological status in the metaphorical conceptual system?
Do “deliberate” metaphors exist independently of “nondeliberate” ones?
How best can we characterize them in a multilevel view of metaphor?
How does the multilevel view of metaphor apply to metaphorical idioms?
Does it shed any new light on their nature?
If visual metaphors are the products of our metaphorical conceptual systems,
which I assume to be hierarchically structured, visual metaphors must also
be hierarchically structures. Is this indeed the case?
What is the “best” method to study metaphors out of the many available
ones in the various disciplines that deal with metaphor?
In what follows, I address the four questions and attempt to examine them in
light of the multilevel view of metaphor.
Deep versus Shallow Metaphors

In a 1995 paper, George Lakoff distinguishes two different kinds of metaphor:
“deep” and “superficial.” He explains the basis for the distinction in the
following way: “The metaphors I have discussed so far in the paper have been
both conceptual in nature and deep, in the sense that they are used largely
without being noticed, that they have enormous social consequences, and that
they shape our very understanding of our everyday world. It is important to
contrast such deep conceptual metaphors such as Morality is Strength and The
Nation is a Family with superficial metaphors, which are only of marginal
interest but which often lead analysts astray” (Lakoff 1995).
In other words, conceptual and deep metaphors are used without being
noticed, they have enormous social consequences, and they shape our under-
standing of the world. Given this characterization, we can ask what the
ontological status of deep and superficial metaphors is – especially in relation
to the four-level hierarchy as described in other publications (Kövecses 2015a,
2017b). How do deep and superficial metaphors fit into the schematicity
hierarchy defined by the conceptual structures of image schemas, domains,
frames, and mental spaces, as shown above?
The particular example Lakoff uses to demonstrate superficial metaphors is
that of the safety net. He illustrates the metaphor from a newspaper article:
“Senator Phil Gramm told a college commencement audience that the social
safety net erected by government by the New Deal and the Great Society had
become a “hammock” that is robbing the country of freedom and virtue” (from
Lakoff 1995).
Lakoff offers an interpretation of the metaphor along the lines of his book
Moral Politics (1996). He states that “[t]he tightrope is straight and narrow – a
moral path.” This is based on the well-known conceptual metaphor: moral is
straight. Given the metaphor, walking the tightrope corresponds to working
and falling off to losing your job.
In addition to this interpretation, several other metaphors seem to be equally
possible in understanding the passage. One of these would be the conceptual-
ization of actions in general as motion. Another would take into account the
balancing act of the tightrope walker. Furthermore, we can think of the
movement of the tightrope walker as a life-related metaphor having journey
as its source domain. Finally, we can interpret the situation depicted by the
sentence as a combination of control and life-action metaphors, in which
controlling life amounts to keeping a job. The set of conceptual metaphors
according to this interpretation would thus be:
action is motion
control is physical balance
life is travel
leading a life is journeying
keeping one’s job is keeping one’s balance while walking
along a tightrope
losing one’s job is losing one’s balance while walking along
a tightrope
The last two metaphors are the most specific ones, given the schematicity
hierarchy mentioned above. They occupy the level of mental spaces. These
highly specific metaphors go together with some everyday knowledge that is
carried over from the source to the target:
The tightrope walker needs a safety net so as not to injure himself if
he falls.
People need social support in case they lose their jobs.
The knowledge provides the basis for the social support of working people.
More importantly for the argument here, the conceptual metaphors identi-
fied above (i.e., control is balance, action is motion, etc.) can be
found at four different levels: the levels of image schema, domain, frame, and
mental space in the following way:
Image schema:
action is motion
control is physical balance
Domain:
life is travel
Frame:
leading a life is journeying
Mental space:
keeping one’s job is keeping one’s balance while walking
along a tightrope
losing one’s job is losing one’s balance while walking along
a tightrope
The schematicity hierarchy works in the following way in this case: Actions in
general are metaphorically understood as (self-propelled) motion (image
schema level). Life is a kind of action and travel is a kind of motion (domain
level). It is a major aspect of life that we live (lead) it, which is understood
through journeying (the activity aspect of travel; see Framenet) (frame level).
A part of leading a life is having (or not having) a job, which is understood as
walking along a tightrope (mental spaces level). The motion and balance
source domains are conceptually integrated (blended) at the level of mental
spaces in the image of “walking the tightrope” (on blending, see Fauconnier
and Turner 2002). As can be seen, both the target and source domains are
more specific elaborations of higher-level metaphors, as we go down the
hierarchy.
This organization of the metaphors tells us about what the distinction
between “deep” and “superficial” metaphors actually means. The deep meta-
phors would be the image schema-, domain-, and frame-level ones. The
superficial metaphor would be the one at the level of mental spaces. In other
words, we can reinterpret and reconceptualize Lakoff’s distinction between
deep and superficial metaphors in terms of the schematicity hierarchy shown
above. The deep metaphors are the ontologically most schematic ones, while
the superficial ones are the least schematic. His terms “deep” and “superficial”
are metaphorical construals of the three “higher” (image schema, domain,
frame) and the “lowest” (mental space) levels.
Deliberate versus Nondeliberate Metaphors

Actually, what Lakoff means by superficial metaphors, such as safety net,
corresponds to or at least greatly overlaps with what Gerard Steen (2013) refers
to as “deliberate” metaphors. Steen defines the latter as follows:
Deliberate metaphor affords conscious metaphorical thought, but is not identical with it
(Steen 2011b). Instead, deliberate metaphor can be defined as the instruction for
addressees to adopt an ‘alien’ perspective on some target referent in order to formulate
specific thoughts about that target from the standpoint of that alien perspective (Steen
2008, 2010, 2011a). This is typically achieved by some form of explicit and direct
metaphor, such as simile. (Steen 2013, www.academia.edu/363751/Deliberate_
metaphor_affords_conscious_metaphorical_thought)
He offers the following example for this kind of metaphor:

Imagine your brain as a house filled with lights. Now imagine someone turning off the
lights one by one. That’s what Alzheimer’s disease does. It turns off the lights so that
the flow of ideas, emotions and memories from one room to the next slows and
eventually ceases. And sadly – as anyone who has ever watched a parent, a sibling, a
spouse succumb to the spreading darkness knows – there is no way to stop the lights
from turning off, no way to switch them back on once they’ve grown dim. At least
not yet. (Steen 2013)
Clearly, in this case, the metaphor is not confined to a single word. It is an
extended description of how we should imagine Alzheimer’s disease. As Steen
points out, in comprehending the metaphor, our view of the disease changes;
we adopt a new perspective on it. How can the multilevel view of metaphor
outlined in this chapter enrich this understanding of deliberate metaphors?
First, let us note that the example is based on the notions of light and
darkness – basic schematic perceptual experiences. We understand a number
of related abstract concepts in terms of the image schema of light: ration-
ality, knowledge, intelligence, truth, morality, and maybe
some others. They all have to do with reason, yielding the metaphor reason
is light. They all have a positive evaluation, as opposed to their antonyms,
emotionality, ignorance, stupidity, falsity, immorality, etc., which
are all negatively valued. In other words, we are dealing here with the image
schema metaphors of reason is light and lack of reason is dark.
Second, the bodily experience of light turning into darkness and darkness
turning into light is used to conceptualize the human mind. The metaphor that
elaborates on the reason is light image-schema metaphor is that of
mental functioning, or more simply, thinking is using light.
The thinking is using light metaphor covers a variety of mental activ-
ities, such as imagining, speculating, observing, considering, etc., which are
understood in terms of ways of using light: vision, envision, visualize, observe,
and, etymologically, specere “to look,” considerare “look at closely,” etc.
from Latin (see Sweetser 1990). The general domain within which this meta-
phor functions is that of the mind; hence the mind as body metaphor (see
Johnson 1987), of which thinking is using light is an example.
Third, a specific case of mental functioning, or thinking, is understanding.
A well-known conceptual metaphor for understanding is: understanding
is seeing (Lakoff and Johnson 1980). This is a frame-level metaphor, where
the specific mental activity of understanding is conceptualized as the specific
sensorimotor activity of seeing – a specific version of making use of light (with
the help of the eyes).
Fourth, the example paragraph for deliberate metaphor spells out the specific
details of the frame-level metaphor. Normal mental functioning corresponds to
having the lights on in a house, whereas its lack corresponds to darkness in the
house. In short, the metaphor at the mental-space level is: lack of normal
mental functioning/lack of understanding (i.e., having alz-
heimer’s disease) is not having the lights on in a house.
If this analysis is on the right track, we can suggest that deliberate metaphors
come with a large nondeliberate part at the image-schematic, domain, and
frame levels. The deliberate metaphor that is presented at the least schematic,
most experience-rich level (the level of mental spaces) is accompanied by and
actually assumes a set of systematic “higher-level” (in Lakoff’s sense,
“deeper”) conceptual metaphors and image schemas.
Another example that comes from my own work (Kövecses 2015a) and
points in the same direction can be offered. An article from USA TODAY is
about cyclist Lance Armstrong’s confessions concerning his doping, and tells
us that his confessions up to that point had not been sufficient to redeem
himself and clean up the sport of cycling. Several (crisis) experts who were
interviewed thought that additional steps must be taken by Armstrong to
achieve this. One specialist in crisis management said this in an interview:
“To use an analogy from the Tour de France, he’s still in the mountain stage,
and will be for some time” (2013, USA TODAY, 6W Sports, Weekly Inter-
national Edition, my italics, ZK).
What we have here is that the specialist has extensive knowledge about
the topic of the discourse, which is Armstrong’s doping scandal. That
knowledge includes that as a cyclist Armstrong participated in several Tour
de France events and that this race has several “mountain stages.” I suggested
that the topic of the discourse primed the speaker to choose a metaphor to
express a particular idea; namely, that, in order to come completely clean,
Armstrong has a long and difficult way to go. This idea was expressed by the
metaphorical linguistic expression in the mountain stage – a deliberate
metaphor. The metaphor is based on the mapping “impediment to motion
➔ difficulty of action (making full confession and being forgiven)” in the
action is self-propelled motion conceptual metaphor. But this is
insufficient as an account for the meaning of the metaphor because there is
still a wide conceptual gap between the meaning of the expression in the
mountain stage and the action is self-propelled motion conceptual
metaphor. The gap can be bridged if we assume the proposed multilevel
conceptual structures in which conceptual metaphors occur in a schematicity
hierarchy, such as the following:
Image schema:
Activities in general are metaphorically conceptualized as entities
moving. This yields the image-schema level metaphor:
ACTIVITY IS
Figure 4.7 Activity is movement
This way of displaying the image-schema level metaphor is used to show the
analogue nature of metaphors at this level. The figure in source domain
position indicates an entity (square) moving (arrow).
Domain:
At the domain level we have the domains of communication, or
communicative action, and self-propelled (forward)
motion. They form the conceptual metaphor:
communicative action is self-propelled (forward) motion
Frame:
There are two frames participating in the conceptual metaphor at the
frame level: confession and race. The metaphor thus becomes:
confessions are races
Mental space:
At the mental space level, the conceptual metaphor is specified as a
bicycle race: confessions are bicycle races. What gets
focused on at this level is the length and difficulty of
confessing wrongdoings. Correspondingly, in the source
concept at this level we have length and difficulty of
being in the mountain stage of the tour de france.
This results in the conceptual metaphor:
the length and difficulty of confessing wrongdoings is the
length and difficulty of being in the mountain stage of
the tour de france
The metaphor is thus fully motivated – both conceptually (i.e., via being a part
of a schematicity hierarchy) and contextually (see Chapter 5). The two
examples I have briefly looked at above indicate not only that deliberate
metaphors have a large nondeliberate part at the image-schematic, domain,
and frame levels. They also show that the deliberate metaphors are actually
based on metaphors that are nondeliberate, and the deliberate and nondeliberate
parts are inseparable, they work jointly within the same schematicity hierarchy
of metaphorical concepts. Moreover, it follows that the metaphors that are
termed deliberate are really both deliberate and nondeliberate at the same time.
One could argue, of course, that some deliberate metaphors are truly isolated
on the mental-space level; that is, unlike the previous examples, they do not
come with a large background conceptual package. Let us consider a potential
example for this possibility. We can take an example used by Steen, sonnet
18 by Shakespeare:
Shall I compare thee to a summer’s day?
Thou art more lovely and more temperate:
Rough winds do shake the darling buds of May,
And summer’s lease hath all too short a date;
Sometime too hot the eye of heaven shines,
And often is his gold complexion dimmed;
One could say that in the sonnet there is a single, isolated deliberate metaphor:
“Shall I compare thee to a summer’s day?” and that there are no deep metaphors
that would unconsciously guide the way we understand and appreciate this part
of the poem. However, a closer look reveals that the various layers of metaphors
that characterize the other examples can be found here as well.
Image schema:
intensity is heat / cold
Domain:
emotion is temperature: intensity of emotion is degree of
heat / cold
Frame:
love is fire: love’s intensity is the degree of heat of fire
Mental space:
the intensity of the poet’s love is the degree of a summer
day’s heat
The heat-cold image schema is well-known in the cognitive linguistic

literature in the conceptualization of emotions (see, e.g., Kövecses 2000b). It
can be found in such conceptual metaphors as love is (heat of ) fire,
anger is (heat of ) fire, affection is warmth, friendly is warm,
unfriendly is cold, etc. In the sonnet, Shakespeare turns the heat of fire
metaphor into a very specific heat-related metaphor: the kind of heat we feel in a
summer’s day (but then he argues against it). Actually, the emergence of and the
motivation for these metaphors is even more complicated if we consider how
our bodily experiences of body heat as metonymy turn into heat of nature (as
from fire and the sun) as metaphor. (On this, see Chapter 3 and Kövecses 2013.)
But the main point of the discussion above is that deliberate metaphors come
with a large nondeliberate conceptual package: the conceptual metaphors that
are presupposed by the most specific level of metaphorical conceptualization –
at the level of mental spaces.
Metaphorical Idioms
As a third illustration of how we can put to use the multilevel view of
metaphor, let us consider metaphorical idioms. It is well-known that particular
idioms can be underlain by conceptual metaphors (e.g., add fuel to the fire –
anger is fire). However, to think of metaphorical idioms in this way would
greatly oversimplify the picture. It seems that metaphorical idioms can be
assigned to very specific metaphorical ideas at one level, but this level is just
one layer of a complex hierarchy of schematicity. Moreover, the hierarchy can
end in a number of image-schema metaphors that jointly characterize the idiom
in question. Finally, for example, idioms, like idioms of money, can belong to
very different hierarchies with very different image-schema, domain-, and
frame-level metaphors (see Kövecses 2018, 2019).
Consider an example sentence taken from a dictionary: Everybody
borrowed money from Joe, the cash cow of the family. Here the metaphorical
expression is cash cow. Let us see what is involved in the metaphorical
conceptualization of the expression and the sentence in the multilevel view
of metaphor presented here.
At the most specific level, that of mental spaces, the person is identified as a
cow and the person’s money as milk. The money comes from the udder of the
cow, so that borrowing money amounts to getting milk from the udder. The
constitutive parts of the compound noun are cash and cow which are conceptu-
ally blended. Blending occurs at the mental spaces level. The dynamic mental
operation of conceptual integration can only take place at this level in working
memory. In sum, on this analysis, the expression cash cow assumes the
following very specific conceptual metaphors (corresponding to a set map-
pings between the source and the target) at the level of mental spaces:
joe is a cow
the money is milk
the source of the money is the udder of a cow
borrowing money from joe is getting milk from the udder of a cow
At the frame level, the last constituent metaphor becomes less specific: Joe
identified as a cash cow at the level of mental spaces is conceptualized as a
source-container and the milk as an object/substance (liquid). In addition, in
the prototypical borrowing-lending frame there is a borrower and a lender. The
borrower requests the money and the lender gives the money to the borrower.
Thus, we get:
borrowing and lending money is requesting the motion of an
object/substance (liquid) from a source(-container) and
the source(-container) causing the motion of the object/
substance (liquid) to the requester
At the domain level, the borrowing-lending frame turns into, or is an

instance of, the more schematic concept of change of possession, which is
conceptualized as the motion of a liquid substance. The person who has the
money is a container and the person who is the source of the money is a
source-container. Thus, the metaphors (or mappings) are:
the person who has the money is a container
the money is a liquid
the person who is the source of the money is a source-container
change of possession of money from person a to person b is motion
(of object/substance (liquid)) from a source-container to a
goal-container
At the most schematic level of image schemas, we have extremely general
metaphorical conceptualizations as a result of the application of a number of
image schemas to the constituent elements of the money domain, such as
persons, resources, sources, occurrences, and changes. These are metaphoric-
ally structured by such image schemas as container, substance, movement, and
motion, yielding the metaphors:
a person is a container
resources are substances
the source of resources is a container
occurrences are movements
change is motion
The point of the previous analysis of the expression cash cow is to show that
we can notice a gradual shift in the schematicity of the conceptual metaphor
that links the various levels. borrowing money from joe is getting
milk from the udder of a cow is more specific than borrowing
money from someone is getting liquid from source-container,
which is more specific than change of possession is motion of sub-
stance out of source-container, which, in turn, is more specific than
occurrences are movements and change is motion. The last two
metaphors are part of the system because we see the borrowing of money from
Joe also as an occurrence and change (of state), and not only as elaborating on
the borrowing-lending frame and the change of possession domain. It is also
significant to point out that the bodily basis of the last two metaphors (i.e., the
correlations between occurrences and movement, on the one hand, and
between change and motion, on the other) makes the entire hierarchical system
immediately meaningful and natural – all the way down to “getting the milk
from the udder.”
We can represent this metaphorical system below as follows. The conceptual
metaphors in boldface are the ones that are related to each other in a schematicity
hierarchy, and they serve as the “backbone” of this metaphor system.
IS:
a person is a container
resources are substances
the source of resources is a container
occurrences are movements
change is motion
DM:
a person having money is a money-container
money is a liquid
the source of money is a liquid-container
change of possession is motion of substance out of source
FM:
the person who is requested to lend money is a source-
container
the money is a liquid-substance
borrowing and lending money is requesting the motion of
an object/substance (liquid) from a source(-container)
and the source(-container) causing the motion of the
object/substance (liquid) to the requester
MS:
the person is a cow
money is milk
the source of money is the udder of a cow
borrowing money from joe is getting milk from the
udder (and the cow giving it)
The description above of the metaphorical idiom cash cow in the given frame
(borrowing) shows the potential conceptual complexity of a single idiom.
Other idioms of money (e.g., cashflow) would require a different schemati-
city hierarchy due to the fact that they would involve other frames (such as
buying and selling). Still different hierarchies would be required by other
money-related idioms, such as “money talks,” where different image sche-
mas would apply (such as force). All these complexities could be fruitfully
studied by making use of the multilevel view of conceptual metaphors as
proposed in the book.
4.6 Visual Metaphors

In the section, I briefly examine how the multilevel view of metaphor applies
to the study of a single visual metaphor – a painting by Mark Tansey, a
contemporary American artist (see Figure 4.8).
4.6 Visual Metaphors 81
Figure 4.8 Mark Tansey: Picasso and Braque (1992).

(Tansey, Mark (b. 1949): Picasso and Braque, 1992. Los Angeles (CA), Los Angeles
County Museum of Art (LACMA). Oil on canvas. 80 × 10800 (203.2 × 274.32 cm).
Modern and Contemporary Art Council Fund (AC1992.154.1). © 2019. Digital Image
Museum Associates/LACMA/Art Resource NY/Scala, Florence)
To begin, in the multilevel view we can work with conceptual structures

(image schema, domain, frame, mental space) that occupy different levels of
schematicity. By setting up such a schematicity hierarchy of conceptual
structures, we can observe that a particular area of experience (such as cubism)
is commonly characterized by conceptual metaphors that involve these con-
ceptual structures on four levels of schematicity. I propose that two aspects of
cubism lend themselves to analyses along these lines; namely, first, that
cubism can be regarded as a complex abstract system and, second, as an
abstract movement. I turn to the metaphor analyses of these two aspects of
cubism below.
Complex Abstract System

In this subsection, I briefly look at cubism as a complex abstract system (see
Kövecses 2010a, 2017b) on the image schema, domain, frame, and mental
space levels.
The notion of cubism, and more generally, art, as represented in the painting
by Tansey can be interpreted as a complex abstract system. Complex abstract
systems are routinely conceptualized metaphorically as complex physical
objects of various kinds. This yields the following highly schematic conceptual
metaphor for the cubism hierarchy: a complex abstract system is a
complex physical object.
At the image schema level, this metaphor can be represented in an analogue
fashion (i.e., here, visually) in the diagram below:
COMPLEX ABSTRACT SYSTEM
COMPLEX PHYSICAL OBJECT
Figure 4.9 Complex abstract system is complex physical object

In the diagram, complex physical objects are represented schematically as a
square. This is the source concept of the high-level conceptual metaphor.
Complex physical objects are characterized by a variety of image schemas,
such as container, link, part-whole, vertical-horizontal
extension, (force), (motion), and so on. These schemas account for
the “complexity” of the physical object. (This complexity is represented by the
grid-like square in the image schema above.)
At the domain level, according to the painting, cubism is metaphorically
viewed as a vehicle, which gives us the conceptual metaphor: cubism is a
vehicle. The source domain of vehicle is conceptually connected to a
number of concepts, such as the following:
creation, shape, size, type, structure, constituent materials,
function, operation, force, motion
Vehicles are created, they have shapes (but not a uniform shape), they have
size and type, they are structured, they are made of some materials, they have
function, they have particular modes of operation, they work with various
degrees of force, and they can move. In short, the concept of vehicle comes
with an extensive domain matrix in Langacker’s (1987) sense.
At the frame level, we have cubist painting as target. The source concept
here emerges as an airplane – a special case of the concept of vehicle. The
aspects of the source as defined by the domain matrix attaching to the concept
of vehicle become more specific here. For instance, airplanes are made in
certain ways, they have a characteristic shape, they are typically large, their
parts are assembled to form a unique structure, they are made of special
materials, and they can fly. These features apply to cubist painting and,
together, they result in the frame-level metaphor cubist painting is an
airplane.
What we know about these aspects individually and taken together greatly
contributes to the metaphorical understanding of cubism and cubist painting.
For example, our knowledge about the type, structure, and constituent mater-
ials of airplanes figures importantly in how we interpret the success of cubist
painting. In other words, to do so, we will employ a submetaphor of the
cubist painting is an airplane metaphor, which is the success of
cubist painting is the appropriate type, structure, and mater-
ials of the airplane.
This takes us to the level of mental spaces. What we find here is how the
potential frame-level knowledge is actually utilized in the painting. First, the
type of the airplane: what we can see is that the frame-level airplane is a very
light airplane. Second, the structure of the airplane: the airplane in the painting
looks like an early one from the experimental stages of airplane building, and it
looks like one which is quite “loosely” put together. Third, the materials: it is
made of cubist paintings (“The Violin” by Picasso).
These features are conceptually blended in the painting. Together they
portray an airplane that is hazardous for purposes of travel. The blend at the
mental spaces level seems to be coherent with Picasso’s and Braque’s initial
fears concerning their project. That is to say, at the level of mental spaces we
are dealing with a very specific conceptual metaphor that can be put as follows:
the fragility and uncertain fate of cubism at the time of its
emergence is the fragility and uncertainty that comes
from the lightness and constituent materials of the air-
plane that is about to take off.
The mental space level of the painting involves more. It also tries to answer the
question of who the leading figure of cubist painting is. It does so by making
use of the “operation” aspect of airplanes. We have the frame level knowledge
that the controller of the plane is the pilot. This maps onto the leading figure of
cubist painting, as represented in the Tansey painting: the pilot corresponds to
the leading figure. Thus we get a further mental space level metaphor: the
leading figure of cubism is the pilot of the plane (that is,
Picasso). Finally at the mental space level, the flying of an airplane frame is
also elaborated at this level: There is a second person (that is, Braque) on the
ground who is actively involved in the operation of the plane (running along
excitedly on the ground, thus possibly playing a less important role in the
operation of the plane). This reaffirms the person of the leading figure: Picasso
is the pilot and leading figure, while Braque is on the ground, figuring less
importantly in the movement.
Abstract Movement
But cubism is not only a complex abstract system, it is also an abstract
movement. Consider the following quote on the relevant sense of “movement”
from a dictionary: “b: a series of organized activities working toward an
objective; also: an organized effort to promote or attain an end - the civil rights
movement - a movement to increase the minimum wage” (from Merriam-
Webster Dictionary, www.merriam-webster.com/dictionary/movement).
At the image schema level the complex abstract system moves and moves
toward a goal. This gives us two image schemas in light of which we interpret
complex abstract systems, such as cubism, as moving: the movement
(activity) and source-path-goal (objective/purpose) schemas,
which can be represented as follows:
Figure 4.10 The movement image schema
Figure 4.11 The source-path-goal image schema
The two image schemas provide two image schematic metaphors: change
is motion and action is self-propelled motion.
More specifically, one level lower we find the following partial character-
ization of the domain of Cubism in Wikipedia: “Cubism is an early-20th-
century art movement which brought European painting and sculpture
historically forward toward 20th century Modern art. Cubism in its various
forms inspired related movements in literature and architecture. Cubism has
been considered to be among the most influential art movements of the 20th
century.[1][2] The term is broadly used in association with a wide variety of
art produced in Paris (Montmartre, Montparnasse and Puteaux) during the
1910s and throughout the 1920s” (from Wikipedia, https://en.wikipedia.org/
wiki/Cubism).
Given the metaphor action is self-propelled motion from the Event
Structure metaphor (Lakoff 1993), we get an additional conceptual metaphor
that applies to cubism in particular:
purposeful artistic activities by a group of cubist artists is
organized travel by a group of travelers
At the frame level, the movement of cubist painting is an air journey,
resulting in the conceptual metaphor cubist painting is an air journey /
a flight. Given this frame level metaphor, the mappings can be laid out as
follows:
the air journey / flight ➔ the cubist painting movement
flying ➔ working / activities within the framework of cubism
the traveler in the air journey ➔ the painter in cubist painting
the destination of the air journey ➔ the goal of cubist painting
the beginning of the air journey ➔ the beginning of the cubist
painting
the end of the air journey ➔ the end of the cubist painting
These are fairly conventional mappings that are involved in any application of
the air journey source concept (more precisely, frame), not just cubism.
However, at this point we can ask why this mode of travel was (unconsciously)
chosen by the artist from among other modes of travel (such as car, train or sea
journey).
There are several answers to this question at the level of mental spaces. One
reason is that the success, viability – the functionality – of the movement was
not clear at the time of its inception. Success and functionality is metaphoric-
ally viewed as upward oriented; hence the metaphor functional is up, as
indicated by the idiomatic phrases up and running, get off the ground, take off.
Now, since action is motion, we can combine it with functional is up
to yield functional/successful action is upward motion, as in the
case of an air journey. Second, we can surmise that the choice is also facilitated
because the initial part of the journey is crucial in air travel (see, e.g., the many
metaphorical expressions concerned with the initial part of air travel: get off the
ground, take off, would not fly, launch). There is clear focus on the initial phase
of the movement in the painting; that is, the mapping that is emphasized is “the
beginning of the air journey ➔ the beginning of the cubist movement.” The
early history of the cubist movement was fraught with difficulties. The other
modes of travel would not have lent themselves to conceptually expressing this
uncertainty and excitement. Third, there might also be some cultural motiv-
ation for this choice, as we’ll see later.
The idea of cubism as an (abstract) movement manifests itself at the mental
space level in several further specific ways. A new, unconventional mapping is
introduced to the frame level metaphor. It is “the initiator / originator of the air
journey ➔ the initiator / originator of cubist painting.” In other words, there is
some focus on the initiators or originators of cubism: namely, Picasso and
Braque. This means that the originators are specified by name, whereas in the
frame level metaphor we only have the role of an unspecified traveler. Actu-
ally, in the painting one of the originators (Picasso) is also the traveler. Thus,
the roles of traveler and originator coincide in the painting (i.e., at the mental
space level). Moreover, in the painting one of the originators (Braque) is
depicted as remaining on the ground, thus receiving secondary status in the
movement. Finally, the painting also evokes the Wright brothers, who had
completed their first successful flight (in 1903) a few years before Cubism as a
movement actually started. The two artists, Picasso and Braque, thought of
themselves as the Wright brothers (Picasso often jokingly called Braque
Wilbur – the first name of one of the Wright brothers). The parallel brings
up for Picasso and Braque, and probably also for Tansey and the viewers of the
painting, the initial excitement associated with the two enterprises. In other
words, blending occurs at this level: the Wright brothers are conceptually fused
with Picasso and Braque.
This reading arises from historical knowledge – in the terminology of the
contextualist view of CMT, the conceptual-cognitive context. The recognition
of this parallel may have also functioned as a motivation for the choice of air
journey as a frame to conceptualize cubism by Mark Tansey.
Methodology
There are a number of different approaches to the study of metaphor within a
general CMT framework. These different methods sometimes compete with
one another and claim superior status in the field (on this, see Gibbs 2017b).
Lakoff and Johnson’s (1980, 1999) work was later seen as an “intuitive”
approach to the study of metaphor. The label was used because Lakoff and
Johnson and their followers were believed to rely on their own intuitions in
identifying metaphors and grouping them into sets of examples to demonstrate
the existence of conceptual metaphors. Corpus linguists argue that to find all or
most of the conceptual metaphors in a language, one needs to take advantage
of large corpora, such as the BNC (British National Corpus) or COCA (Corpus
of Contemporary American English) (see, e.g., Charteris-Black 2004; Deignan
2005; Stefanowitsch 2006). They emphatically point out that our linguistic
intuitions are not adequate for studying metaphors. More recently, Kövecses
and his co-workers (Kövecses 2015b; Kövecses et al. 2019) proposed to revive
the older intuitive method. Following Kövecses’ (1986) initial suggestion,
I call the revised method the “lexical approach” and show that the lexical
method has advantages that the corpus-linguistic (more precisely, corpus-
based) method cannot ignore. Other metaphor scholars place emphasis on
the social-pragmatic uses of particular metaphorical expressions. They typic-
ally work with smaller databases than corpus linguists, and study the commu-
nicative functions of metaphors in authentic discourses (e.g., Cameron 2003;
Musolff 2006; Semino 2008). Some scholars pay attention to how frames and
metaphors are related and study various metaphorical constructions (see, e.g.,
Sullivan 2013). They often make use of the Framenet project that was
developed by Charles Fillmore.
Psycholinguists and cognitive psychologists study metaphors as a concep-
tual phenomenon and the embodied nature of metaphorical conceptualization
(see, e.g., Boroditsky 2001; Casasanto 2009; Gibbs 1994, 2006; Gibbs and
Colston 2012). This work is largely experimental – in the form of either “in
vitro” or “in vivo” experiments (see Kövecses 2005, on this distinction).
Neuroscientists take advantage of the most recent brain imaging techniques
and try to identify the neuronal activities that underlie the use of conceptual
metaphors in the brain (e.g., Bambini et al. 2011; Coulson 2008; Gallese and
Lakoff 2005 for an overview). Finally, groups of scientists are working to
build computational models of how humans use conceptual metaphors for a
variety of purposes (see, e.g., Feldman 2006; Narayanan 1999).
Thus, we have at least the following approaches dedicated to the study of
conceptual metaphors:
Intuitive approach
Corpus linguistic approach
Lexical approach
Discourse analytic approach
Framenet type approach
Psycholinguistic experimentation
Neuroscientific experimentation
Computational modeling
There are no doubt other approaches, but even this set shows very clearly the
variety of methods for the study of conceptual metaphor.
How can we pair off the different approaches with the different levels of
metaphor described above? The following pairing of approaches with the
levels is suggestive, rather than definitive. Most of the approaches can be used
to study several different levels. I suggest that the particular approaches apply
primarily to particular levels of metaphor, as shown in Figure 4.12.
IMAGE SCHEMAS – Psycholinguistics; Neuroscience
--------------------------------------------------------------------------------------------------------
DOMAINS
– Intuitive approach; Lexical approach; Framenet
FRAMES
– Discourse analysis; Computational modeling
ME NTAL SPACES – Corpus linguistics
– Psycholinguistics; Neuroscience
Figure 4.12 Levels of schematicity and methods of metaphor research
What lends significance to these pairings of levels and methods is threefold:

First, there is no single approach that can be used to study all levels of
metaphor. Second, several distinct approaches can be used to study the same
level(s), but they can contribute complementary insights to it/them. Third,
certain approaches may be better fitted to study a particular level than others.
To demonstrate the first point in the previous paragraph, I will briefly look at
the “lexical method” and corpus linguistics. Researchers using the lexical
method search for various lexical items or other types of information that are
related to the general topic, or concept, under investigation (such as particular
emotions indicated by particular lexemes: e.g., anger, fear, surprise). These
include synonyms, antonyms, related words, various idioms and phrases,
collocations, and, importantly, even the definitions of the lexemes (see
Kövecses 2015b; Kövecses et al. 2019). The most likely sources for these
types of information are dictionaries: monolingual and bilingual dictionaries,
thesauri, collocation dictionaries, idiom dictionaries of various sorts, and, in
general, any collections of words and phrases related to a concept. Very
importantly, researchers prefer and tend to use dictionaries that offer example
sentences and usage notes (about register, frequency, provenance, etc.) for the
linguistic expressions (words, idioms, collocations, etc.) that they contain for
the lexeme in question. Since it is not possible to find all of this information in
a single dictionary, several different dictionaries of various kinds must be
consulted before one can collect (possibly) all the lexical information that
characterizes a lexeme under investigation.
The various lexical items that belong to a particular concept, or domain, are
dominantly types, not tokens. The linguistic expressions that are identified by
the dictionaries represent lexemes. In this method (that is based on mostly
dictionary data), it is not possible to gage the actual frequency of the tokens,
which is a drawback of the method. Furthermore, since the lexical method

deals with types, that is, with linguistic expressions that have become lexica-
lized, the types provided by the dictionaries represent the most conventional-
ized linguistic expressions of a language related to a domain. The types that are
identified are at the image schema, domain, or frame levels, that is, the level of
decontextualized language (that Kövecses 2002/2010 calls the “subindividual”
and “supraindividual” level, respectively), as opposed to the mental-spaces
level, that of contextualized linguistic usage (of the types) by individual
speakers who use tokens. (This level was referred to by Kövecses 2002/2010
as the “individual” level.)
We should note here that corpus linguistics, as another method, was pro-
posed in the study of language, including metaphor research, to correct the
deficiencies of work at the supraindividual level; specifically, to draw attention
to contextualized linguistic usage in real discourse, i.e., the individual level
(see, for example, Deignan 2005; Semino 2008; Stefanowitsch 2006) and to be
able to do quantitative analyses of tokens. Both of these are legitimate pro-
posals. But they work at a level that is different from the level where the lexical
approach operates. The lexical approach works with types attaching to frames
and domains, while various corpus methodologies work with tokens (as well as
types) at the level of mental spaces.
The lexical approach can uncover the most conventionalized metaphorical
linguistic expressions related to a target domain (i.e., the types) on the basis of
which researchers can hypothesize the existence of systematic conceptual corres-
pondences between two domains – the conceptual metaphors. The conceptual
metaphors can be assumed to be shared by speakers of a language who have
acquired the metaphorical expression types relating to a target. Experimental
studies indicate that this is indeed the case (see, e.g., Gibbs 1994, 2006; Gibbs
and Colston 2012). Thus the lexical approach can reveal a considerable portion of
the shared metaphorical conceptual system in a linguistic community, though not
the metaphorical conceptual systems that particular individual speakers possess.
By contrast, corpus linguists suggest that the lexical approach does not work
for the individual level of metaphor analysis. This criticism is valid. But it
should be kept in mind that the lexical approach was not designed to capture
contextual variation in the use of metaphors and the individually variable
metaphorical conceptual systems of individual speakers (on these issues, see
Kövecses 2005, 2010, 2015b). There are several other more specific criticisms
coming from corpus linguists (see, e.g., Kövecses 2011a, 2015b), but the
discussion so far suffices to draw attention to the issue of relationship between
the multilevel view of metaphor and the various methodologies that are
employed in metaphor studies. Most important of these, in my view, is the
following: The different levels of metaphor are associated with different goals
and these goals go together with different methods. If we want to characterize
how individual speakers use metaphors in context, we use corpus- and

discourse-based approaches; if we are interested in making hypotheses con-
cerning our (largely) shared metaphorical conceptual system, we can legitim-
ately employ the lexical approach.
4.7 Conclusions
In the chapter I offered a four-level view of metaphor, which was demonstrated
with the building source concept. The suggestion was that the four levels
form a hierarchy in which conceptual metaphors occupy different levels of
schematicity, the highest level being that of image schemas and the lowest that
of mental spaces, with domains and frames in between. In the view proposed,
each conceptual metaphor is characterized by all four levels.
The mappings between the source and target concepts occur on the same
level: domains correspond to domains, frames to frames, and mental spaces to
mental spaces. In the case of the level of image schemas, the image schemas
structure highly schematic target concepts, such as complex abstract
systems, whose members are, for instance, the concepts of theory, mind,
nation, company, life, career, economy, and many others. The map-
pings at the various levels license metaphorically used words at different levels
(such as inside (image schema level), structure (domain level), and builder
(frame level), as shown above.
The four levels are connected vertically, in that the lower levels are more
specific versions of the higher, more schematic levels. Schematicity was
characterized, following Langacker (1987), as “relative precision of specifica-
tion” that applies to a concept.
Thus, conceptual metaphors occur at four levels of schematicity in an
interlocking vertical hierarchy of image schemas, domains, frames, and mental
spaces. In other words, conceptual metaphors cannot and should not be linked
to a single conceptual structure, such as frames or domains. Conceptual
metaphors are complexes of all four of these at the same time. I call this
structuring the “multilevel view” of metaphor.
The schematicity hierarchy is characterized by two major divisions. One is the
distinction between image schemas, on the one hand, and the other three levels,
on the other. While image schemas are analogue conceptual structures, the other
three are non-analogue ones (i.e., propositional) in nature. The other division
that was proposed is between levels that, on the one hand, belong to long-term
memory (image schemas, domains, frames), as opposed to the level of mental
spaces, that function online in working memory. Mental spaces activate higher-
level conceptual structures, while the latter structure mental spaces.
Insofar as the lexical approach to the study of metaphor can show, the
building source concept highlights two aspects in relation to the target
4.7 Conclusions 91
concept of complex abstract systems: the aspects of construction

and logical support, which are captured by the building (as process)
and the (physical) support frames (including strength of support). We
can call these aspects the “main meaning foci” of the building source
domain – aspects of a concept that are conventionally or habitually profiled
in metaphorical conceptualization. However, we also saw cases where aspects
of building outside the main meaning foci were profiled (through the parts
of a building and function of a building frames). These non-routine
projections require more complex mechanisms to account for their use (as we
saw, e.g., in the fire-exit example).
The constructs that I call “main meaning foci” are the profiled aspects of
culturally imbued complex conceptual metaphors. They seem to be closely
linked to the primary metaphors that make up complex ones in the sense that
the main meaning foci reveal which primary metaphors constitute a complex
metaphor. It was noted, however, that the notion of the main meaning focus
applies more broadly than primary metaphors, in that they extend beyond
correlation metaphors.
I suggested that the metaphors used at the least schematic level, i.e., that of
mental spaces (or scenarios), take on pragmatic functions and are used in ways
that do not characterize metaphors at higher levels of schematicity. These
differences can be explained with reference to the difference between the
functioning of long-term memory as opposed to working memory. As the
example of the caretaker indicates, several of the features that characterize
the level of mental spaces can jointly apply to a metaphor used at this level.
Finally, the multilevel view of metaphor was shown to have several impli-
cations for conceptual metaphor theory. Of these, four were discussed in the
chapter. First, concerning the issue of “deep” versus “superficial” metaphors
and their ontological status, it can be suggested that “deep” versus “superficial”
metaphors can now be given a more precise ontological status. Deep meta-
phors are those that are at the frame, domain or schema levels. Superficial ones
are those that function at the level of mental spaces, that is, typically uncon-
ventional metaphors in discourse that can activate a number of conceptual
metaphors at all of the “higher” levels. Second, regarding “deliberate” meta-
phors, I proposed that deliberate metaphors can be reconceptualized as mani-
festing entire schematicity hierarchies of concepts. That is, deliberate
metaphors, as we saw, come with a large nondeliberate part. They evoke
frame-, domain-, and schema-level metaphors. Given this, the metaphors that
are taken to be deliberate are also nondeliberate in their conceptual back-
ground. In the view presented here, the two kinds of metaphors (deliberate
and nondeliberate) are inseparable. Third, I showed by way of an example how
schematicity hierarchies can apply to metaphorical idioms. The hierarchies can
reveal not only new dimensions of conceptual complexity in the analysis of
idioms but can also shed new light on how many idioms are organized around
a large “core” of high-level metaphors. Fourth, as regards how the multilevel
view of CMT applies to the analysis of visual metaphors, I examined a single
painting by American artist Mark Tansey from the perspective of extended
CMT. The multilevel view of metaphor proved especially helpful in account-
ing for the interpretation of the painting. Both the painter and the viewers
approach the work at a variety of levels of schematicity. Much of the concep-
tual materials on the image schema, domain, and frame levels are most
probably shared by the artist and the viewers. They are, to a large extent,
conventional and stored in long-term memory. It is on the mental space level
that the artist’s creative imagination can be captured. In a way, it can be
suggested that the painting is the expression of the conceptual product of the
artist’s relevant mental space in the course of producing the painting – without
the metaphorical processes that led to the final (static) product (i.e., the
painting). These are, however, recoverable from the product itself. Fifth and
finally, I discussed the issue of methodological “rivalry” in relation to the
question of what the “best” method is in the study of metaphors. Here, the
suggestion was that different methodologies and approaches apply to different
levels of metaphor and that they are all useful in doing their respective jobs.
All in all, it seems that by applying the multilevel view of metaphor we can
gain a great deal of coherence and uniformity in the description of seemingly
diverse phenomena in the study of metaphor.
5 Conceptual or Contextual?
Conceptual metaphor theory has been primarily formulated as a cognitive

theory – lacking a contextual component. Consider now how the theory would
run into trouble if it is conceived as a mainly or exclusively cognitive phe-
nomenon. Here is what it would look like: Conceptual metaphors are defined
as sets of mappings between a concrete source domain and an abstract target
domain. The mappings are conceptual correspondences between elements of
the source and those of the target. Linguistic expressions make these corres-
pondences manifest in language, and the meaning of the expressions is based
on the particular mappings. By their nature, the systematic mappings form a
part of the conceptual system. The central cases of these metaphors arise,
according to the standard version of the theory, from correlations in experience
between abstract (subjective) and sensorimotor concepts, that is, from
embodied experience.
However, if we accept this view as the only or as our complete account of
metaphor, it becomes very difficult to explain the comprehension and produc-
tion of many metaphors – both conceptual and linguistic – in particular
situations of discourse (see, e.g., Kövecses 2010b, 2015a). As we full well
know, metaphor comprehension is highly context-sensitive, as experimentally
demonstrated by decades of psycholinguistic and psychological research. (For
overviews, see Gibbs 1994, 2006; Gibbs and Colston 2012.) The production
aspect of metaphor use has received hardly any or, in any case, a lot less
attention over the years in CMT. I have shown in some recent publications (see
Kövecses 2010a,b, 2015a) that the use of many naturally occurring linguistic
and conceptual metaphors is virtually impossible to explain without taking into
account the role of context. In other words, CMT needs to embrace a context-
ual component – in addition to its other components, such as its conceptual
structure, cognitive function and embeddedness, embodied character, and
neural basis. While all of these are essential, they do not help us account for
why and how speakers/conceptualizers produce particular linguistic and con-
ceptual metaphors in specific discourse situations. In this chapter, based on my
previous work, I identify and outline some of the most important elements that
should be part of such a contextual component of CMT. In doing this, I also
93
follow the Langackerian tradition in cognitive linguistics, where meaning

(both literal and figurative) is always dependent on context (see Langacker
1987). More recently, Langacker formulates this foundational tenet in the
following way: “The prior discourse is a major determinant (along with
context, background knowledge, etc.) of what I call the current discourse
space (CDS). The CDS is a mental space comprising everything presumed to
be shared by the speaker and hearer as the basis for discourse at a given
moment” (Langacker 2008: 59).
In a way, I consider it my task in this chapter to make explicit what is
contained in this mental space when a speaker engages in metaphorical
conceptualization in a specific discourse situation (see also Kövecses 2015a).
In other words, my primary focus will be on the role of context in the
production of metaphors.
A contextualist version of conceptual metaphor theory requires the charac-
terization of at least three large issues that center around the following three
questions: (1) What are the elements of (metaphorical) meaning making?; (2)
What are the most common contextual factors that play a role in the use and
creation of metaphors?; and (3) What is the cognitive mechanism through
which contextual factors actually produce metaphors in natural discourse? In
the remainder of the chapter, I draw heavily on my 2015 book Where Meta-
phors Come From.
5.1 Elements of Meaning Making

First of all, I am assuming that the prototype of meaning making (including
metaphorical meaning making) is a situation in which speaker/conceptualizer 1
and hearer/conceptualizer 2 are co-present and share a great deal of infor-
mation in and about the situation, including themselves. Obviously not all
communicative situations are like this, but when they are not, it usually
weakens the potential to understand each other.
In other words, meaning making in general depends on the notion of
“relevant context,” which means the information that the participants find
especially relevant for a meaningful exchange. The question of the relevant
context is usually formulated from the hearer/conceptualizer 2’s perspective
(see, e.g., Sperber and Wilson 1986/1995). However, from the production side
of metaphors the question for the speaker/conceptulizer1 becomes: How can
I make use of the present context to create the appropriate metaphor that a
listener can understand and interpret correctly? But Sperber and Wilson also
offer the notion of “mutual cognitive environment,” which seems to be
compatible with the idea of context as developed here. I will outline the most
common components of such a cognitive environment (i.e., contextual factors)
in a later section.
5.2 Types of Context and Kinds of Contextual Factors 95
In a communicative situation, a speaker performs three actions by means of

one (Sinha 2007): (1) representing a referential situation, (2) expressing his or
her communicative intention, and (3), directing the listener’s “intentional
processes toward the referential situation” (Sinha 2007: 1282). Thus, a sphere
of joint attention is created between the speaker and the hearer.
In discourse that employs metaphors, the use of metaphorical language
functions as joint action, and joint action requires common ground between
speaker and hearer. Using symbols is a form of “joint action” (Clark 1996),
which assumes a large amount of shared knowledge between speaker and
hearer called “common ground” by Clark (1996). For Clark (1996), the shared
basis for the common ground can be of two kinds: “communal common
ground” and “personal common ground.” A “cultural community” is a group
of people who share a great deal of information that others don’t (Clark 1996:
101), the information being stored in frames. People share most of the dis-
tinctive frames that are based on conventional conceptual metaphors. There is
also “personal common ground,” which derives from shared perceptual experi-
ence and joint action. This can also be used as a metaphorical source domain in
metaphorical communication. Such cases of metaphor use occur frequently,
say, when journalists report on an event and they use metaphors that are based
on some distinctive perceptual property of the event.
5.2 Types of Context and Kinds of Contextual Factors

In Where Metaphors Come From, using examples from naturally occurring
discourse, I have presented a large amount of evidence to show that the use of
metaphors in discourse is influenced by a variety of contextual factors. These
contextual factors can be grouped into four large categories: situational con-
text, discourse context, conceptual-cognitive context, and bodily context.
Before I summarize and discuss the various context types and the contextual
factors they subsume, I introduce two general constraints on metaphor use.
The use of metaphors in discourse is extremely variable. This is because
context plays a crucial role in which metaphor is used in which situation. Each
and every communicative situation is different, and this results in variation in
metaphorical conceptualization. We can distinguish two large sets of factors
that seem to play a role in metaphor variation: those that have to do with
differential experience and those that have to do with differential “cognitive
styles.” The contextual factors presented below have to do with differential
experience. They consist of some contentful knowledge (i.e., one that has
conceptual content) that reflects (direct or indirect) experiences of the world.
These are experiences that can trigger the use of particular metaphors. In order
for conceptualizers to produce and comprehend metaphors, they need to have
access (directly or indirectly) to the experiences that are made use of in the
metaphors. These experiences provide the common ground that allows con-
ceptualizers to produce and comprehend contextually induced, or generated,
metaphors in discourse.
The factors under cognitive styles, by contrast, reflect particular ways in
which experiences of the world need to be presented, given the prevalent
cognitive conventions and preferences of a language community. Such issues
as at which level of generality a metaphorical idea is presented (schematicity),
how it should be framed (framing), to what degree it should be conventional-
ized (conventionalization), which aspect of the body it should involve (experi-
ential focus), and others, are presentational in nature (for example, see
Kövecses 2015a). In general, the former set of factors (i.e., differential experi-
ence) respond to the question of “what” can prompt or prime the use of certain
metaphors, whereas the latter set to the question of “how” metaphorical
conceptualization needs to be presented in a language community. The latter
factors listed under differential cognitive style function as constraints on the
speaker-conceptualizer only, who is to follow the linguistic and cognitive
conventions of the language community. The hearer (conceptualizer 2/com-
prehender) is not constrained in this way.
Now let us turn to the various types of context and the more specific
contextual factors that constitute them. (For a detailed discussion and a number
of examples, see Kövecses 2015a).
Situational Context
The situational context comprises a variety of contextual factors. Most
commonly this type of context can be thought of as including the physical
environment, the social situation, and the cultural situation.
The physical environment can shape metaphorical meaning making. It
includes the flora, the fauna, the landscape, the temperature, the weather,
perceptual properties of the situation, and so on. For example, it is a common
observation that American English metaphors relating to the physical environ-
ment are characteristically different from those of other English-speaking
countries (see, e.g., Kövecses 2000c). The small-scale, local environment,
such as the visible events in or the perceptual properties of a situation, can
also make its influence felt in shaping metaphors.
The social situation consists of social aspects of life that typically center
around notions such as gender, class, politeness, work, education, social organ-
izations, social structure, and others. All of these can play a role in metaphorical
conceptualization. For example, Kolodny (1975, 1984) shows that American
men and women developed very different metaphorical images for what they
conceived of as the frontier in America in the period before the twentieth
century. While the women commonly thought of the American frontier as a
“garden to be cultivated,” men conceptualized it as “virgin land to be taken.”

(For several additional examples on gender metaphors, see Kövecses 2005.)
Gender can of course also be thought of as part of culture, not only a social
issue. In general, the distinctions in this section serve heuristic purposes only.
The cultural situation involves both the global context (the shared know-
ledge represented in the conceptual system) and the local context (the specific
knowledge in a given communicative situation). An example of how the
global context can affect metaphorical conceptualization is the way different
concepts can produce differential metaphors in different cultures and lan-
guages, such as the metaphors for anger: anger is heat (of fluid or
solid) in a large number of languages such as English and Hungarian,
whereas in Chinese the metaphor can also involve gas as its source domain –
as a result of the influence of Yin and Yang theory (see Yu 1998). The more
immediate local context can play a similar role in the production of metaphors
(see Kövecses 2010b).
Discourse Context
The discourse context involves the surrounding discourse, knowledge about
the main elements of discourse, the previous discourses on the same topic, and
the dominant forms of discourse related to a particular subject matter.
The surrounding discourse is simply the linguistic context – often referred
to as “cotext.” Viewed from the perspective of the discourse’s producer (the
speaker), elements of the preceding discourse (either by the speaker/concep-
tualizer 1 or the hearer/conceptualizer 2) can influence the (unconscious)
choice of metaphors, as was shown in an example taken from The Times by
Kövecses (2010b): “which helped to tilt the balance – and Mr Hain – over the
edge.” In this case, the contextually induced metaphor arises from the elliptical
use of the verb tilt in the phrase tilt Mr. Hain over the edge. It is the presence of
the word tilt in the immediate cotext that leads to the second use of the
metaphor.
Conceptualizers often rely on their knowledge concerning the main elements
of a discourse: the speaker, hearer, and the topic. For example, it occurs
commonly that in many newspaper articles knowledge about the topic as a
contextual factor leads to the creation of new metaphors (Kövecses 2010b,
2015a). An example of this involved David Beckham in an article in which the
journalist remarked: “Los Angeles Galaxy are sardines not sharks in the ocean
of footy” (Kövecses 2010b). Here the relevant knowledge includes that Beck-
ham played for the Los Angeles Galaxy soccer team and that Los Angeles is
located on the ocean with all kinds of fish in it.
The metaphors in one discourse can also derive from previous discourses on
the same topic. This can take a variety of forms ranging from elaborating,
extending, questioning, negating, reflecting on, ridiculing, to otherwise taking

advantage of a metaphor previously introduced. For example, an MP in the
British Parliament responded to the then Prime Minister Tony Blair who said
he does not have a reverse gear (i.e., he can only go forward) with the
following statement: “but when you’re on the edge of a cliff it is good to have
a reverse gear” (example taken from Semino 2008). This was a humorous
twist induced by the prior discourse on the progress is motion forward
conceptual metaphor.
Certain forms of discourse can acquire dominant status in a community.
When a dominant form of discourse related to a subject matter emerges, the
metaphor used in or based on this discourse can become widespread both
temporally (historically) and spatially (cross-culturally). For example, the
discourse of Christianity commonly gives rise to the use of metaphors in the
Christian world.
Conceptual-Cognitive Context
This type of context includes the metaphorical conceptual system, ideology,
knowledge about past events, and interests and concerns.
Concepts can stand in a metaphorical relationship with one another (e.g.,
life is a journey, argument is war) in long-term memory. Given such
metaphorical relationships between concepts (such as between, say, argu-
ment and war), their presence or absence in the metaphorical conceptual
system may lead to the production and comprehension of particular metaphors.
(If the metaphorical connection does not exist in the user’s conceptual system,
he/she would not be able to come to use a metaphor.) A metaphorical concep-
tual system can function as context in this sense. Given an intended metaphor-
ical meaning (say, “supporting an argument”), we can search the conventional
metaphorical conceptual system for the best choice of metaphor. This happens
in cases where a conventionalized metaphorical meaning (“supporting an
argument”) is expressed via a conventional linguistic metaphor (such as
defend), with a matching target element activating the relevant mapping, or
conceptual correspondence, in an existing conceptual metaphor (e.g., the mean-
ing “supporting an argument” by means of the word defend in the argument
is war conceptual metaphor as based on the mapping “defending one’s physical
position in war corresponding to “supporting one’s position in an argument”).
The conventional metaphorical conceptual system is especially utilized this way
when there are no other overriding contextual factors in the discourse situation
that might influence the (unconscious) choice of a metaphor.
Ideology can also be a formative factor in how metaphors are used in
discourse. One’s ideological stance concerning major social and political
issues may govern the choice of metaphors (as work by, for instance, Goatly
2007, shows). A good example of how ideology might influence the choice of
metaphors is George Lakoff’s (1996) study of American politics, where
conservatives tend to use the nation is a strict father family meta-
phor, while liberals prefer the nation is a nurturant parent family
version of the generic metaphor the nation is a family. For another
example, we can mention the Marxist version of the society is a building
metaphor with talk about “superstructure,” and so on, or the Marxist idea of
“class struggle.” Goatly (2007) is an important exploration into the metaphor-
based ideology of capitalism.
Being aware of past events and states (i.e., items in short-term and long-term
memory) shared by the conceptualizers may also lead to the emergence of
specific metaphors in discourse. A special case of this involves a situation in
which the speaker assumes that the hearer has a particular mental state. Such
memories of events can belong to the life of a community or an individual. It has
been often observed that the memory of historical events can lead to the produc-
tion (and comprehension) of some metaphors (see, e.g., Deignan 2003, Kövecses
2005). Different historical contexts can create differential preferences for par-
ticular life metaphors among Hungarians and Americans (see Kövecses
2005). The particular events in a specific communicative situation preceding
an act of metaphorical conceptualization may also produce similar effects.
People are commonly prompted to use particular metaphors (more pre-
cisely, metaphorical source domains) relative to their interests and concerns
about the world (see Kövecses 2005). Entire groups and individuals can be
said to have certain characteristic interests or concerns that may affect the way
they make meaning metaphorically. Since Americans are commonly regarded
as dynamically oriented, rather than passive, in their attitude to life, and,
relatedly, are sports-loving in general, it is not surprising that they use a large
number of sports metaphors. Similarly, if a person has some kind of profes-
sional interest, that person is likely to draw metaphors from his or her sphere
of interest (see Kövecses 2005). In other words, the kinds of activities we
engage in routinely affect the metaphors (the source domains) we use (see
Gelfand and McCusker 2001).
Bodily Context
A particular state of the body can produce particular metaphorical conceptual-
izations in specific cases, such as a poet’s or writer’s illness. Elsewhere,
I showed how Dickinson’s choice of metaphors may have been influenced
by her optical illness (see Kövecses 2010b). Such cases illustrate, more
broadly, that people’s bodily specificities influence which metaphors they tend
to use. There is experimental evidence (Casasanto 2009) that left-handers
prefer to use the moral is left, as opposed to the moral is right
conceptual metaphor. Such metaphors contrast with the metaphors that evolve
on the basis of the general properties of the human body (i.e., the correlation-
based primary metaphors – see Chapter 1).
Given this kind of evidence and the many metaphorical examples that are
based on local and temporary specificities of the human body (i.e., not on
universal embodiment), one might take the body as a further form of the
context. In this view, the body – especially those aspects of it that are activated
in the ongoing communicative situation – can influence the choice of meta-
phors in natural discourse (see Kövecses 2015a). The body is not only respon-
sible for the production of hundreds of conceptual metaphors through the
many correlations between subjective and sensory-motor experience (cf.
Grady, 1997a,b; Lakoff and Johnson 1999), but it can also prime the use of
particular metaphors in more immediate, local contexts (as further experimen-
tal evidence by, e.g, Boroditsky, 2001; Boroditsky and Ramscar, 2002; Gibbs
2006; Gibbs and Colston 2012 indicates). In other words, the body can lead to
the production of metaphors in discourse in the same way as the other
contextual factors previously mentioned can. Given this, we can think of it
as an additional context type.
5.3 Local and Global Context

Within the varied set of contextual factors that were briefly introduced above,
two general types of context can be distinguished: local and global (see
Figure 5.1). The local context involves the specific knowledge conceptualizers
have about some aspect of the immediate communicative situation. Thus, the
local context implies specific knowledge that attaches to the conceptualizers in
a specific communicative situation. It corresponds, at least roughly, to Clark’s
personal common ground. By contrast, the global context consists of the
conceptualizers’ general knowledge concerning their community’s environ-
ment (physical, social, cultural). It involves knowledge shared by an entire
community of conceptualizers. Thus, it is close to Clark’s communal common
ground. The distinction local and global context is mostly of theoretical nature.
In many actual communicative situations, there is no sharp dividing line
between the two types of context.
5.4 Operationalizing Contextual Influence on Metaphor Production

In the present section, I attempt to briefly discuss the main conceptual ingredi-
ents that appear to be prerequisite for the production of a particular metaphor-
ical expression in a discourse situation. I propose that there are four cognitive
processes involved in this (without claiming that they are used in a conscious
way): (1) Out of the many construal, or cognitive, operations the speaker must
5.4 Operationalizing Contextual Influence on Metaphor Production 101
L CO
OBA NT
EX
GL T
CON
CAL TE
situational LO XT discourse
context context
metaphor
used
conceptual-cognitive bodily
context context
Figure 5.1 Summary of context types

(from Kövecses 2015a)
decide on metaphor as a meaning-making device – as opposed to some other

device. (2) Given the various context types, the speaker is exposed to a wide
array of information or experiential knowledge that “compete” for the
speaker’s attention. (3) One or several of the contextual factors will emerge
as the strongest and prime(s) the speaker to use the matching metaphorical
expression. (4) The expression will have the meaning that is appropriate for the
communicative intention of the speaker and it results from the conceptualizer’s
ability to build a legitimate “conceptual pathway.” I label the required cogni-
tive processes as follows:
Deciding on cognitive operations
Being exposed to various forms of local vs. global context
Being primed by contextual factor(s)
Building a conceptual pathway
Deciding on Cognitive Operations

In any given context, the speaker unconsciously chooses a particular cognitive,
or construal, operation (potentially supported by others) out of the many that are
available to him/her. These include schematization, figure-ground alignment,
scalar adjustment, dynamic and static attention, viewpoint selection, subjective
or objective construal, metaphoric or metonymic conceptualization, conceptual
integration, and others (based on Langacker 2008; see also Kövecses 2006).
Often, several of these constitute alternative choices for the construal of a
situation. The alternative operations commonly include abstraction, schematiza-
tion, attention, perspective (subjectivity-objectivity), metonymy, metaphor, and
conceptual integration. In the case under consideration, the speaker (uncon-
sciously) decides on using metaphoric conceptualization.
Being Exposed to Various Forms of (Local or Global) Context

The diagram below summarizes the various context types, as discussed so far:
The conceptualizers are exposed to a large amount of shared information in
a communicative situation. As explained above, the global context involves
knowledge that members of a language community share. The local context
consists of knowledge that individuals have in and about a particular ongoing
communicative situation. The shared meaning making system provides a
(more or less) uniformly present context for all members of a language
community, while the latter involves specific local aspects of one’s under-
standing of a specific communicative situation. The various context types and
contextual factors can be seen as “competing” for the attention of the con-
ceptualizers – factors on which metaphors may be built.
Priming
It can be argued that (the knowledge and awareness of ) our experiences in the
local and global contexts can prompt the use of particular metaphors – either
conventional or novel ones. The various contextual factors can prime the use
of a metaphor in discourse. Priming is a well-studied cognitive process used
extensively in psychological and psycholinguistic experiments with a sizeable
literature (see, e.g., Boroditsky and Ramscar 2002; Casasanto 2009; Gibbs and
Colston 2012; and several other studies). Importantly, priming is based on the
simulation of some experience in the situational, discourse, bodily, and
conceptual-cognitive context. It is shared experience (the dynamically evolv-
ing common ground in a situation) that enables the production and comprehen-
sion of metaphors in discourse.
5.4 Operationalizing Contextual Influence on Metaphor Production 103
Let us take an example for this from work by Semino (2008), as described
by Kövecses (2015a). Semino studied the metaphors used by various partici-
pants at the 2005 G8 summit meeting in Scotland on the basis of an article
about the summit. The summit was accompanied by a major rock concert
called Live 8. Some participants assessed what the G8 summit had achieved
positively, while some had doubts concerning its results. Semino looked at one
such negative assessment she found in an article about the summit. She
comments in the following way:
In contrast, a representative of an anti-poverty group is quoted as negatively assessing
the G8 summit in comparison with the Live 8 concert via a metaphor to do with sound:
1.1 Dr Kumi Naidoo, from the anti-poverty lobby group G-Cap, said after “the roar”
produced by Live 8, the G8 had uttered “a whisper.”
The reference to “roar” could be a nonmetaphorical description of the sound made by
the crowd at the concert. However, the use of “whisper” in relation to the summit is
clearly a (negative) metaphorical description of the outcome of the discussions in terms
of a sound characterized by lack of loudness. Hence, the contrast in loudness between
the sounds indicated by “roar” and “whisper” is used metaphorically to establish a
contrast between the strength of feeling and commitment expressed by the concert
audiences and the lack of resolve and effectiveness shown by the G8 leaders. (Semino
2008: 3–4)
In my view, the metaphor whisper here emerges largely from the physical
(-social) context in which it is produced. Dr. Kumi Naidoo creates the meta-
phor whisper against a background in which there is a very loud concert and a
comparatively quiet summit meeting. We can think of the loudness and the
relative quiet of the occasion as perceptual features of the two events. Thus, the
original conceptualizer, Dr. Kumi Naidoo, chooses a perceptual property of
the physical context from all the experiential content that is available to him.
Conceptual Pathways
Out of the many potentially metaphor-producing experiences represented by
the various contextual factors above, a particular metaphorical meaning can be
conveyed if a specific target-domain meaning is expressible by a context-
induced source-domain meaning. This can be achieved if a particular concep-
tual pathway (made up of several conceptual metaphors and metonymies) can
be built between the two meanings (Kövecses 2015a). This way a particular
piece and kind of information (or experiential content) and, with it, a particular
context-induced metaphor is chosen out of the huge number of available
options in the situation.
A conceptual pathway is often a schematicity hierarchy, as discussed in
Chapter 4. The safety net, the in the mountain stage, and other examples show
how a contextual meaning can be created and made sense of by a particular

schematicity hierarchy. I will discuss this aspect of the proposed view in more
detail in the following chapters.
In conclusion, we can suggest that in actual communicative situations
speakers/conceptualizers derive their metaphors from four types of experi-
ence: the situational, discourse, conceptual-cognitive, and bodily contexts.
This goes against the traditional belief that most of our linguistic metaphors
are simply conventionalized linguistic expressions that have a certain
meaning, and we use the metaphors when we wish to express those
meanings in a given context. It also goes against a commonly held view
in cognitive linguistics that the metaphors we use are simply based on
conceptual metaphors in our heads. Instead, the four context types and the
contextual factors belonging to them prime conceptualizers to (uncon-
sciously) choose their metaphors in discourse. More specifically, the
priming effect can take place only if the conceptualizers (both speaker
and hearer) can build the appropriate conceptual pathway between the
intended target-domain meaning and the particular experiential content
that is primed.
5.5 Contextual Influence and Levels of Schematicity

The last remaining question I wish to deal with here is this: At which level of
schematicity does the context make its influence felt on the (unconscious)
choice of conceptual metaphors as outlined in Chapter 4? If we think of
conceptual metaphors as multilevel structures, this question becomes
inevitable.
The source domains are not only coherent with their targets (or the other
way around), but also with context in a broad sense. Elsewhere, I proposed that
many metaphorical source domains are coherent with one or more contextual
factors. This constraint was called the “pressure of coherence” with the context
(see Kövecses 2005, 2015a). There seems to be abundant evidence for this
kind of activation in the experiments conducted by Boroditsky, Casasanto,
Gibbs, Colston and other researchers (see, e.g., Boroditsky 2001; Casasanto
2009; Gibbs 2006; Gibbs and Colston 2012). But, again, the question remains
which one of the levels is affected – no matter how obvious the answer
might be.
Since conceptualizers are aware of and share contextual factors of various
kinds and types in the course of online communication, the level of schema-
ticity involved is that of mental spaces, which are present in the course the
functioning of working memory in a specific communicative situation. This is
so even if the mental spaces that are created activate various frames, domains,
5.6 Metaphorical Creativity: The Role of Schematicity and Context 105
and image schemas. With the creation of each new mental space, the context
changes dynamically, and it affects how a situation is conceptualized meta-
phorically in and through that mental space.
At the same time, of course, the repeated, long-term effect of a contextual
factor can lead to the (re)construction of a frame or domain that bears the
impact of that frame or domain (Figure 5.2). In other words, in a sense, frames
and domains are also context-dependent (not only mental spaces), but their
shaping by context occurs over an extended period of time (both in terms of
content and structure), unlike the online priming of mental spaces, which is
instantaneous.
All in all, then, we have the following picture:
analogue
structure
IMAGE SCHEMAS
…………………………………………………………………………
long-term memory
DOMAINS
non-analogue structure
FRAMES
------------------------------------------------------------------------------------
MENTAL SPACES
working memory
CONTEXT
Figure 5.2 The influence of context on schematicity hierarchies
The upward arrow indicates that it is at the level of mental spaces that
contextual influence on metaphorical conceptualization occurs.
5.6 A Note on Metaphorical Creativity: The Role of Schematicity

and Context
How does this view of the joint workings of levels of schematicity, on the one
hand, and context, on the other, affect the way we think about metaphorical
creativity? An important aspect of metaphorical creativity is elaboration. We
can think of elaboration in two ways. One is where a concept, like journey,
is realized lexically in two ways (see, e.g., Lakoff and Turner 1989). But there
is another way of thinking about elaboration. In it, conceptual metaphors may
be related to one another at different levels of schematicity. An example is
provided by the American poet Karl Sandburg. Consider the first stanza of the
poem called Skyscraper:
BY day the skyscraper looms in the smoke and sun and

has a soul.
Prairie and valley, streets of the city, pour people into
it and they mingle among its twenty floors and are
poured out again back to the streets, prairies and
valleys.
It is the men and women, boys and girls so poured in and
out all day that give the building a soul of dreams
and thoughts and memories.
(Dumped in the sea or fixed in a desert, who would care
for the building or speak its name or ask a policeman
the way to it?)
The metaphoric-metonymic use of the skyscraper is clearly prompted by the
situational (more precisely, physical-cultural) context. The poem was written
in 1916 in Chicago, and it was at the turn of the twentieth century that
skyscrapers began to be built on a large scale in major American cities,
including Chicago.
Here the image of the skyscraper is based on the metaphor society is a
building and this conceptual metaphor is part of a schematicity hierarchy.
At the level of image schemas, we conceptualize complex abstract systems
(such as societies) as complex physical objects. At the domain level, societies
are conceptualized as buildings and the creation of a society as the physical
creation of a building. The latter is the main focus of the metaphor in the poem.
At the frame level, the idea of creation is elaborated as construction, which
results in the conceptualization of the construction of a society as the construc-
tion (i.e., building) of a building (with tools and ingredients). Finally, at the
mental spaces level, the building of a new American society is conceptualized as
building a skyscraper (with hammers and crowbars and spikes and girders).
Thus, we get a set of conceptual metaphors in a schematicity hierarchy of
conceptual metaphors:
Level of Image schema:
complex abstract systems are complex physical objects
Level of Domain:
society is a building; the creation of a society is the physical
creation of a building
Level of Frame:
the construction of a society is the building of a building
(with tools and ingredients)
Level of Mental spaces:
building a new american society is building a skyscraper
(with hammers and crowbars and spikes and girders)
5.7 Context and Visual Metaphor 107
The two highest levels, those of image schemas and domains, may be universal
(or at least widespread) in cultures when it comes to the metaphorical concep-
tualization of societies. But the frame level, and especially the lowest mental
spaces level are fairly culture- and even individual-specific. Sandburg captures
the idea of building a new American society through the idea of building a
skyscraper. At this level of conceptual metaphor, other poets may diverge
widely (though using the same conceptual metaphors at the higher levels),
depending on a large number of contextual factors (such as the physical
environment). In other words, we can conclude on the basis of this example
that elaborations of high-level conceptual metaphors may contribute to meta-
phor variation in poetry as well. Metaphors that may be universal or wide-
spread at higher levels of schematicity may turn into something that is specific
to particular poetic traditions or to individual poets.
The example demonstrates very clearly the usefulness of both the notion of
schematicity hierarchy and that of contextual influence in the production of
novel conceptual metaphors.
5.7 Context and Visual Metaphor

In this section, I propose to discuss the following issue in the study of visual
metaphors (see, e.g., Cienki and Müller 2008; Forceville 1996, 2008; Force-
ville and Urios-Aparisi 2009). It concerns the issue of what should be (or can
be) considered a visual metaphor in relation to the interplay between context
and visual metaphor.
By visual metaphors I mean “pictorial metaphors” in the sense of Forceville
(2016), who offers the term on the basis of his study of static pictures and
objects. Following in his footsteps, El Refaie defines pictorial metaphors as
follows: “. . . I will use the term ‘pictorial metaphor’ to refer to images of
concrete, recognizable objects, scenes, and people that are used to stand for
something else” (El Refaie 2019: 16). (El Refaie additionally includes spatial
and stylistic metaphors in the category of visual metaphor.)
We can begin by observing that visual metaphors come in a large variety of
forms (see, e.g., Forceville 2016). In dealing with some problematic issues in
the theory of visual metaphors, Forceville usefully outlines several ways in
which visual metaphors may be distinguished: First, he classifies visual meta-
phors according to how they are formed (contextual metaphors, hybrid meta-
phors, simile, verbo-pictorial metaphors, and integrated metaphors). Second,
he distinguishes visual metaphors according to whether the source is inside or
outside the story in which it is presented (diegetic vs. non-diegetic metaphors).
Third, visual metaphors vary according to genre; they can be advertisements,
cartoons, comic books, operas, theatre plays, films, gestures, traffic signs,
drawings, pictures, paintings, illustrations, statues, and many others. This variety
of visual metaphors is based on a genre-based classification of visual experience,
in that the forms of visual metaphors belong to different genres as convention-
ally distinguished. In them, we find conceptual metaphors that have been
discovered mostly on the basis of linguistic metaphors but also ones that are
specific to particular genres (see, e.g., El Refaie 2019). Fourth, the classification
of visual metaphors may also be based on their degree of conventionality. They
can be creative or structural (i.e., nonconventional or conventional). Fifth, visual
metaphors can be classified according to whether they employ a single modality
or several ones (monomodal vs. multimodal metaphors).
There are, however, other ways of classifying visual metaphors. Another
potential basis for the classification can be the degree of the sign-like nature of
the visual experience that constitutes the visual metaphors. My contention is
that visual metaphors can be composed not only of signs (i.e., sign-like visual
experiences), such as in advertisements, cartoons, and sculptures, but also of
visual experiences that we would not normally consider sign-like, such as
when we see a building or look at a scene in nature.
The suggestion builds on the idea that non-sign-like visual experiences
constituting visual metaphors utilize the same processing mechanisms as the
sign-like ones. The interpretation and creation of both kinds of metaphorical
visual experience relies on framing, mapping, recognizing correlations between
experiences, finding resemblance between experiences, creating new experiences
represented by new frames, getting primed by experience, resorting to already
existing metaphors in long-term memory, and several others. In other words,
conceptual metaphors realized visually are not to be found outside the human
organism; they are found internally as human beings process experiences, either
sign-like or non-sign-like, by means of these conceptual mechanisms.
5.8 Two Kinds of Visual Metaphor

In the section, I discuss two kinds of visual metaphors: visual metaphors that
are correlation-based and those that are resemblance-based. (On the distinction
between correlation and resemblance metaphors, see Grady 1999.) My object-
ive is to show that non-sign-like visual experiences can also give rise to visual
metaphors in the sense that such experiences can become source domains of
visual metaphors in the same way as sign-like experiences can.
Correlation-Based Visual Metaphors in the Real World

Let us begin with a conceptual metaphor that has a large number of visual
manifestations: control is up (Lakoff and Johnson 1980). The metaphor
5.8 Two Kinds of Visual Metaphor 109
manifests itself in several linguistic expressions, including “be on top of the

situation,” “be the underdog,” “have control over someone,” and “climb to
the top” (Figure 5.3). This is a primary metaphor (see Grady 1997a,b) that
is commonly exemplified visually by, among other things, the winner’s
rostrum:
Figure 5.3 Winner’s podium

(https://pixabay.com)
Here, the winner of a competition occupies the highest position at the award
ceremony because he or she has social control over the other competitors. That
is, the person with the most power or control is placed highest. Social power or
control is reflected in higher physical position.
There are many other social situations where we can see the conceptual
metaphor at work. For example, the king sits or stands higher than his subor-
dinates, the speaker of the house sits higher than the representatives, teachers
often stand on a podium in classrooms, and so on. In general, elevated
platforms of various sorts, besides their practical functions, serve the purpose
of indicating how some people have more power than others.
In many other cases it may not be immediately obvious that the metaphor
control is up is present and it metaphorically constitutes certain aspects of
reality (Figure 5.4). By way of an example, imagine that you are a foreign
tourist in a city and you encounter the following visual image:
Figure 5.4 Buda Castle

The image is about Budapest and it shows the Buda Castle, the former Royal
Palace, on top of the castle hill on the Buda side of the river Danube, which
separates the older part of the city, Buda, from the newer part, Pest.
The image we can see above is the photograph of a “real” visual image that
a foreign tourist (or, as a matter of fact, any other person) can see in Budapest.
Is the real image (perceived by a tourist) and/or the photograph of the image
(shown above) in any sense metaphorical? And, importantly, how can we
find out?
It should be noticed that the image does not portray people who could be
said to be in some obvious power-based relationship with one another. We can
only see buildings, a hill, the river, boats, etc. And yet, interestingly, foreign
visitors to Budapest can interpret what they see via the control is up
metaphor.
In a small and informal research project, I investigated the metaphors that
foreign tourists who visit Budapest use to interpret their experiences (see
Kövecses in press). I did a simple Google search to find descriptions of
people’s experiences and impressions of Budapest. I only gathered descriptive
statements that contained metaphorical expressions. This way, a database of
about 130 pages long was compiled. After eliminating duplicates and the
repeated uses of the same metaphor, I ended up with an 83-page-long list of
metaphorical statements in their natural discourse environment about Buda-

pest. Most of the statements came from individual tourists who have visited the
city in recent years. They left their notes about Budapest in English, but
English is not necessarily their native language. I did not try to separate the
comments by native speakers of English from those by nonnative speakers.
It seems to me, though, that most of them are native speakers. It is important to
mention that a certain percentage of such statements came from tourist propa-
ganda (e.g., from travel agencies), and are thus not spontaneous expressions of
people’s opinion about the city. I did not try to filter out this kind of material
from my database either. In general, then, it can be said that I did not perform a
formal corpus-linguistic data collection and analysis. One of the major sights
for visitors to Budapest is the Buda Castle. In their comments on their experi-
ences, the foreign visitors often use descriptions such as the ones below:
Buda Castle is certainly a tourist favorite with it’s [sic] mighty position,
elevated high above the city.
Buda is high up on the hill and has a lot of upper class neighborhoods, while
Pest is where the younger crowd tends to live and has a lot more to do. (see
picture below)
It appears that when tourists perceive these views, they think about them in
terms of power; they seem to perceive some controlling force attaching to
them. In other words, they tend to interpret the sight through some well-known
primary metaphors, such as power is up and control is up. In the first
case, the “elevated position” of the castle viewed in correlation to the power
that comes from (and in former times, it actually came from) a building in that
physical position. In the second case, again, the “neighborhoods high up on the
hill” are correlated with the notion of social control (“upper class neighbor-
hoods”), which seems to be based on the control is up metaphor.
How is this metaphorical connection possibly made by looking at this view?
I suggest that it is likely to be made via metonymic thinking. A conventional
metonymy that obtains between a place and the people who live or work in that
place is the place for the people living/working there, or, for
short, place for the people in that place. The castle is not just a
building but also the place where there are powerful people. As a result, the
building can metonymically stand for the people with power. In addition, the
city as a geographical location metonymically stands for the people who live
there. This means that the metonymy the place for the people in that
place can apply twice: once to the castle and once to the city. The castle will
now metonymically stand for powerful people and the city for ordinary people.
Thus, a perceived situation where there were originally no people (in a
relationship of power) will be seen as a power relationship between people.
The double application of the place for the people in that place
Figure 5.5 View of the Danube at Budapest

metonymy will be input to the power/control is up metaphor. This results

in visitors interpreting the situation in terms of power and control – and not just
in terms of a geographic location.
In a similar way, the sight of the Danube is also couched in metaphorical
conceptualization (see Figure 5.5). The Danube, in other words, is not simply a
perceived entity that moves and divides the city into two parts (Buda and Pest).
The comments visitors make suggest that people “see” a lot more than just a
big body of moving water. According to one commentator:
The river is glorious. It is a sight that inspires poets to writing and philoso-
phers to contemplation. The water was coursing with momentum, the sun
was reflecting its rays from the surface, and boats roared through the waters
with grace and power. On the side you see a rocky cliff standing firmly by a
mounted castle. I was mesmerized.
What makes a river glorious (italics in the original)? What is it that inspires
poets and philosophers, as well as tourists? I believe the answer has to do with
the metaphors that the Danube evokes in the people standing by it and watching
it. The expression “rocky cliff standing firmly by a mounted castle” evokes,
again, the metaphor power/control is up, where the castle represents
power and control. But in addition, the sight of the river by itself can activate
in its viewers such metaphors as powerful action is forceful motion
(coursing with momentum, boats roared through the waters, with grace and
power). This metaphor derives from the primary metaphor action is motion.
Moreover, the entire scene also involves the good is light or bright
metaphor, as shown by the sun’s rays reflecting from the surface of the river.
Resemblance-Based Visual Metaphors in the Real World

In the previous examples, aspects of the real world are interpreted metaphoric-
ally and metonymically. The examples are situations in the real world where
the physical and social are correlated, like the winner (social) standing on the
top of the podium (physical). But there are other situations where we find some
kind of resemblance between the physical and the abstract at a very high level
of generality. In such cases, an aspect of the real, physical world can evoke an
abstract idea that shares some skeletal structure with it. When visual metaphors
emerge from such skeletal structures, we are not dealing with primary meta-
phors. The metaphors evoked will be resemblance-based.
Context, including the physical situation, can provide conceptualizers
with visual images that can constitute the source domains of visual meta-
phors. The metaphors can be generated by a conceptualizer being able to
find some very skeletal resemblance between the structure of the visual
image and the structure of some abstract (non-visual) idea. The abstract idea
that shares the skeletal structure with that of the visual image will become
the target domain.
Probably all of us had the experience when a scene in the real world led us to
the recognition of a particular abstract idea because we felt that the scene and
the idea are somehow similar. This can also happen in poetry. My favorite
example for this happening is Matthew Arnold’s poem, Dover Beach (see
Kövecses 2010b, 2015a). In the poem, Arnold is describing a scene in which
he is present (or as if he was present), together with another person:
The sea is calm to-night.
The tide is full, the moon lies fair
Upon the straits,- on the French coast, the light
Gleams and is gone; the cliffs of England stand,
Glimmering and vast, out in the tranquil bay.
Come to the window, sweet is the night-air!
(www.artofeurope.com/arnold/arn1.htm)
This is a more or less literal description of what he can see around him from the
window of a seaside house (Figure 5.6). But when the tide goes out he (they)
may see something like this:
Figure 5.6 Dover Beach

The visual scene consists of low tide at Dover Beach with the sea withdraw-
ing, which gives him the following idea:
The sea of Faith
Was once, too, at the full, and round earth’s shore
Lay like the folds of a bright girdle furled.
But now I only hear
Its melancholy, long, withdrawing roar,
Retreating, to the breath
Of the night-wind, down the vast edges drear
And naked shingles of the world.
(www.artofeurope.com/arnold/arn1.htm)
The stanza may be interpreted in the following way: The sea that covers the shore
gives protection to the earth, but when it withdraws at low tide, it leaves the earth
exposed. The same can happen to faith. Its importance for people may diminish,
but when it does, it leaves people unprotected. In other words, the sea corresponds
to faith, the earth corresponds to the people, and the covering of the earth by the
sea corresponds to the protection of the people by faith. Thus, the visual image of
the low tide can trigger thoughts in the poet, which is based on some skeletal
structure shared by the withdrawing of the tide and the diminishing of faith.
However, this recognition of shared schematic structure can also occur in
reverse order. It may well be that the poet had these ideas in mind about
5.9 Conclusions 115
Christian faith, and then he found the appropriate visual image to express it. As
a matter of fact, in all probability the two can also co-occur simultaneously.
But this is not the crucial point. What is important is that creative, unconven-
tional resemblance metaphors can emerge as a result of visual images of the
real world and can combine with abstract ideas.
In this section, I noted that, in addition to visual aspects of many metaphor-
ical source domains (as in, e.g., building that is found in the metaphor
theories are buildings), there are visual images (e.g., the sight of Buda
Castle or Dover Beach at low tide) in the online physical context of metaphor-
ical conceptualization that can serve as source domains of novel metaphors.
I suggested that we can think of such cases as the visual images priming the
selection of metaphorical source domains.
In Chapter 4 and the present one, I examined a variety of visual experiences
that appear to evoke visual metaphors. This is a range of experience types that
extends from sign-like visual experiences (like a painting) to non-sign-like
visual experiences (like a visual image of an artifact or nature or a city scene).
Visual metaphors are evoked by paintings through winner’s podiums all the
way to cityscapes and scenes in nature. The last two (non-sign-like) cases,
cityscapes and natural scenes, are not commonly subjected to serious examin-
ation from a CMT perspective. However, they provide us with a new challenge
in the study of visual metaphors, in that they greatly extend the range of visual
experiences that might give rise to visual metaphors.
At the same time, remarkably, the comprehension or interpretation of all of
these visual experiences, including sign-like and non-sign-like ones alike,
makes use of the same metaphorical processing mechanisms. The visual
metaphors that are evoked by visual experiences can be based either on
correlations or resemblance. This latter bifurcation applies equally to non-
sign-like visual experiences, such as cityscapes and scenes of nature.
5.9 Conclusions
In the chapter, I argued for a broad conception of context in metaphorical
conceptualization – one that covers our cognitive interaction with various
elements and properties of the situation of discourse, the discourse itself, the
conceptual-cognitive background, and the body of the speaker and hearer.
These dimensions of our interaction with the world were labeled the situational
context, discourse context, conceptual-cognitive context, and the bodily con-
text. All of these context types and the specific contextual factors that make
them up can influence the creation of metaphors in discourse.
This view of metaphor is, I feel, preferable to other views within cognitive
linguistics because it extends the study of metaphor beyond those that are
body-based in the usual sense in conceptual metaphor theory (i.e., universal
correlations between sensorimotor experience and abstract ideas), while at the

same time acknowledging the centrality of body-based metaphors. The meta-
phors based on the situational, the discourse, the conceptual-cognitive context,
together with the bodily one (involving unique features of individual bodies)
may far outnumber universal correlation-based metaphors.
The four context types and the contextual factors belonging to them prime
conceptualizers to (unconsciously) choose their metaphors in discourse. More
specifically, the priming effect can take place only if the conceptualizers (both
speaker and hearer) can build the appropriate conceptual pathway between the
intended target-domain meaning and the particular experiential content that
serves to prime the conceptualizer. There can be multiple context types and
contextual factors that can have a priming effect.
Furthermore, the approach reveals that a comprehensive theory of metaphor
cannot simply stop with providing an explanation of what makes certain
metaphors universal. It also needs to take into account what makes other
metaphors so variable across cultures, within cultures, and even across
individuals.
What makes this last point possible is the extension of conceptual metaphor
theory along contextualist lines. In particular, first, conceptual metaphors and
their linguistic expressions do not only arise from bodily experience. Many
conceptual metaphors and their linguistic manifestations can emerge from
various context types and contextual factors; that is, they are all context-
induced metaphors, and not necessarily based on some previously existing
universal bodily experience. Second, as regards theories of context, if the
suggestions made above are valid, we can have a broader view of context than
is traditionally the case. Most theories of context conceive of context as the
linguistic or the situational environment for the use of a given metaphor. With
the addition of the conceptual-cognitive and bodily context, we can get an
extended view of context. Third, regarding cognition in general, the context-
ualist view of metaphor takes us beyond the idea that embodiment provides the
only grounding for human cognition. The view proposed here seems to
indicate that cognition is grounded in multiple ways, including our situational,
linguistic, bodily, and cognitive-conceptual experiences and knowledge. The
contextual factors represent the totality of our experiences as we interact with
the world. This can provide us with a new framework in the study of meta-
phorical meaning in cognition and culture.
Finally, the visual experience of the physical environment is a crucial aspect
of the physical context, as we saw in the case of the cityscape and Dover Beach
examples. In the case of non-sign-like visual experiences, the visual experi-
ence of the physical environment primes conceptualizers to evoke the appro-
priate conceptual metaphors in the interpretation process.
6 Offline or Online?
One of the major criticisms of conceptual metaphor theory has been that it is
not capable of accounting for meaning in actual occurrences of metaphorical
language in real discourse (see, e.g., Cameron 2003; Deignan 2005; Goatly
1997; Semino 2008; Steen 2007, and many others). This kind of criticism
makes perfect sense if we consider the multilevel view I am outlining in this
book. In the “standard” view of CMT, researchers work on the levels of image
schema, domain, and frame. These are conceptual structures that are decon-
textualized patterns in long-term memory that can account for metaphorical
meaning in the most general ways. On the one hand, image schemas can
explain why certain metaphors can be meaningful through embodiment and,
on the other, the conceptual metaphors on these levels (image schema, domain,
and frame levels) can account for metaphorical meaning through the mappings
that constitute the conceptual metaphors and the inferences that can be made
on the basis of such mappings. However, both of these conceptual mechanisms
(providing meaningfulness and accounting for decontextualized meanings) are
fairly general ones that cannot capture the contextually variable richness of
metaphorical meaning in naturally occurring discourse.
What makes it possible to get at this richer and more specific aspect of
metaphorical meaning, but at the same time preserve these general aspects of
meaning, is to think of metaphor as a multilevel phenomenon that occurs also
at the level of mental spaces. As I have noted in previous chapters, online
metaphorical activity necessarily makes use of the conceptual structure of
mental spaces. This is the level where we produce and comprehend metaphors
in natural discourse. Specifically, it is at this level that speakers
use metaphors online (i.e., produce and comprehend metaphors)
use them in a fully contextualized way
use them with specific socio-pragmatic functions
add emotional value to them
create novel metaphors
use metaphors deliberately
create metaphors as a result of contextual priming
117
use individual metaphors

add lexical elaborations to frames
perform conceptual integration
mix metaphors in the same discourse
and possibly several others.
In this chapter, my main goal is to survey some of the most important online
processes and functions of metaphorical language as related to mental spaces
and point out the connections of this language to conceptual metaphors at the
higher levels.
6.1 Various Uses and Functions of Metaphors

at the Level of Mental Spaces
I will briefly discuss and exemplify some of the uses mentioned above in the
present section. Most of the examples that I use in the section to demonstrate
my points will be taken from the concept of anger, an experiential domain that
is well studied by cognitive linguists, and so sufficiently familiar to many
readers. Later on in the chapter, I will look at two of the uses above in more
detail – those of mixing metaphors and conceptual integration.
Metaphors in Natural Discourse Occur at the Mental Spaces Level

In online communication, speakers mobilize the static image schemas,
domains, and frames at the mental spaces level. Consider the following
example from Google:
“Everything that I wanted to say I couldn’t think how to say in French. I was bursting
with anger, mostly at myself for my lack of French vocabulary.” (my bolding, ZK)
http://vagendamagazine.com/2014/10/harassment-on-the-streets-of-paris-why-is-it-worse-
than-in-the-uk/
In the example, the metaphorical expression “burst with anger” assumes the
metaphorical conceptualization of the speaker’s anger as the high degree of
physical pressure inside the person-container (speaker). We can put this in the
conventionally used metaphor format: the high intensity of the
speaker’s anger is the high degree of the physical pressure
inside the speaker-container. This is a very specific conceptual meta-
phor at the mental space level, in that it fills the role of the angry person with a
particular value (the speaker) and it is built on a specific aspect of anger;
namely, its high intensity. This specific metaphor is based on the angry
person is a pressurized container metaphor and one of its mappings:
the intensity of a person’s anger is the degree of pressure
inside a person. This, in turn, is an elaboration of the more schematic
6.1 Uses and Functions of Metaphors at Mental Spaces 119
metaphor emotions are forces; in particular, on its mapping emotional

intensity is degree of pressure. The image-schematic metaphor that
subsumes these metaphors is intensity is strength (of effect).
Thus, we get the following schematicity hierarchy:
Image schema level:
intensity is strength (of effect)
Domain level:
emotions are forceful interactions: emotional intensity is
degree of pressure
Frame level:
the angry person is a pressurized container: the intensity of
a person’s anger is the degree of physical pressure inside a
person-container
Mental spaces level:
the high intensity of the speaker’s anger is the high degree
of physical pressure inside the speaker-container
This example is simply meant to indicate that the metaphor burst is produced
and comprehended in real discourse at the level of mental spaces, and the
mental space that is most immediately used to produce and comprehend it is
linked systematically to conceptual metaphors at higher levels that jointly
serve the process of the production and comprehension of the metaphorical
expression.
There Is Contextual Priming at This Level

We can suggest that in the previous example contextual priming is effected by
the conceptual-cognitive context, an aspect of which is the metaphorical
conceptual system, where the expression “burst with anger” can be found
attached to a conventional conceptual metaphor. But, as was noted in Chap-
ter 5, the priming effect can come from other context types as well, such as the
discourse context, the situational context, and the bodily context. Let us look at
an example for the first type, where a novel metaphor (hot stove) is introduced
as a result of the contextual influence of discourse:
Ryuuji couldn’t help but shout at Taiga to ease his anger. Even though he was almost
freezing, his anger was like a hot stove inside him. Looking at Taiga, Ryuuji breathed
out large clouds of white mist as he rubbed his numb hands against his cheeks, slowly
easing some blood into his hands. (my italics, ZK, www.baka-tsuki.org/project/index
.php?title=Toradora!:Volume10_Chapter1)
In the passage a person is described as very angry but also freezing. We can
ask how the hot stove metaphor is (unconsciously) chosen in the discourse to
metaphorically render the person’s anger. It can probably be assumed that
mention of the same person’s freezing has a role to play in this. Freezing is the
opposite of body heat that is associated with anger and that gives rise to heat-
related anger metaphors. The image of a hot stove is evoked as a result of the
mental-space level interaction of the being cold frame and the body heat
frame, in that a hot stove can be used to prevent someone from being cold.
A hot stove is a device that can be a part of the being cold frame but also of
the body heat frame that serves as the basis for the anger is heat
conceptual metaphor. In addition, a stove is a container just like the angry
person’s body is a container. In sum, at the level of mental spaces an aspect of
the discourse (literally freezing) can prime the speaker to introduce a non-
conventional metaphorical expression that also matches an established con-
ceptual metaphor (anger is heat).
Blending Occurs at This Level

Fauconnier and Turner have repeatedly claimed (see, e.g., Fauconnier and
Turner 2002, 2008) that blending is a dynamic online process that makes use
of mental spaces. One of their examples to demonstrate this is “He was so mad,
smoke was coming out of his ears”(Fauconnier and Turner 2002; Turner and
Fauconnier 2000). This metaphorical expression can be found in texts of all
kinds, including the one below:
He looked as if smoke was coming out of his ears, a vein about to pop, and his
handsome face red from anger. (www.wattpad.com/336753058-the-luna%27s-alpha-
meet-the-pmsing-alpha-and/page/2)
Before the metaphorical expression became conventionalized in various forms

(such as smoke / steam / fumes coming out), it must have emerged as a blend in
some naturally occurring discourse, such as the above, when a speaker merged
the image of an angry person’s head with the image of a fire with smoke or a
container with steam in it. Actually, it is also possible that the “original” blend
was a cartoon or drawing, and it gave rise to the linguistic metaphor. But even
if this was the case, the person who drew the blend must have created the blend
by making use of mental spaces online as part of a dynamic process. (I will
come back to the discussion of this example and conceptual integration in
general later in the chapter.)
Pragmatic-Narrative Functions Are Given to Metaphors

at This Level
At the image schema, domain, and frame levels, the linguistic metaphors
express a particular aspect of the conceptual content of the source domain.
6.1 Uses and Functions of Metaphors at Mental Spaces 121
The meaning of these metaphorical expressions is based the topology of the

source domain and its mapping onto the target; their meaning is the expression
of this conceptual content (such as the cause of the fire-heat being mapped onto
the cause of one’s anger). But at the mental spaces level (i.e., where the
metaphorical linguistic expressions are used in actual discourse) a lot more
can happen as regards meaning, as indicated by the following example:
“1 Jimmy: Uh: I was (.) boiling at this stage and 2 I was real angry with Connie ( ). And
3 uh went up to bed ‘n (.) I lay on the bed.” (Edwards 1999: 274)
Edwards makes the point that this particular use of “boiling” establishes
accountability and provides justification for one’s (the speaker’s) actions.
The passage is part of an interview in a counseling session, in which Jimmy
talks about how his wife’s “provocations” during an evening in a pub led him
to leave her in the pub (with the man she was “messing” with) and to go home.
The use of “boiling” shows that the intensity of his anger matches the intensity
of the “provocations,” thus making his wife accountable for his anger. At the
same time, it also shows that the intensity of his anger provides justification for
leaving the pub, go home, and, later, even assault his wife. “Boiling anger”
presents him as being passive in relation to events in the sense that he appears
to believe that such a high intensity of anger justifiably leads to those actions
on his part. Clearly, the use of “boiling” here goes far beyond the expression of
some metaphorical content. It fulfills extremely important narrative-discursive
functions that constitute another and indispensable part of the expression’s
meaning in a particular context.
Elaboration for Rhetorical Purposes

Metaphors at higher levels can be elaborated at the level of mental spaces. The
elaborations allow speakers to improvise and be creative. For this reason the
process of elaboration is commonly employed in poetry. A well-known
example in CMT is Adrienne Rich’s poem “Phenomenology of Anger” (see
Gibbs 1994):
“white acetylene / ripples from my body / effortlessly released / perfectly trained / on
the true enemy // raking his body down to the thread / of existence.”
The poem is based on the anger is a hot fluid metaphor. The poet
elaborates the (hot) fluid as acetylene that can ripple from the body. Acetylene
is a dangerous substance that can cause serious injury to the person who is
targeted with it. The poet elaborates on an aspect of the conventional meta-
phorical source frame in an unpredictable and creative way. The process
results in novel metaphors that serve the rhetorical purpose of the author.
Combined Uses at the Level of Mental Spaces

The five examples of the uses of metaphors at the level of mental spaces only
serve demonstrative purposes. As we saw, there are several additional uses of
metaphors at this level. But there is one general observation that I would like to
underscore at this point. It is that, although so far I have tried to keep the
various uses of metaphors in real discourse distinct, in reality they commonly
occur together in particular cases of usage. Consider the following example
mentioned earlier in the book (see Chapter 4), taken from Musolff (2001):
“What does he [Chancellor Kohl] need this house for, after so many years as
Chancellor? – Well, it’s obvious, he wants to become the caretaker.” (Die Zeit,
May 16, 1997)
Here, the want mental space is structured by two frames: by the become
frame (unrelated to the discussion here) and the function of the build-
ing frame, where there is a caretaker as a possible frame element. When the
second speaker employs the term, (s)he is using it in the course of processing
the discourse online up to that point. Also, by using the word caretaker, (s)he
is creating a novel or unconventional metaphor in a specific context. Moreover,
on at least one interpretation, the second speaker expresses some disapproval
regarding the role of a caretaker. The point I am making through this example
is that metaphorical linguistic expressions that make mental spaces manifest
(that is, at the least schematic level of conceptual structures) in real discourse
are used in ways that do not apply to metaphorical linguistic expressions that
make manifest higher level structures. This is because the latter expressions are
stored in long-term memory conventionally denoting some conceptual content,
and they are not used in working memory where all of the various uses and
functions of metaphors in context can occur over and above the expression of
some conceptual content.
In the rest of the chapter, I would like to offer two, more detailed case
studies of online metaphorical activity. One concerns the phenomenon of
mixing metaphors in natural discourse and the other deals with an assessment
of conceptual integration theory vis-à-vis conceptual metaphor theory.
6.2 Mixing Metaphors

How does conceptual metaphor theory handle mixed metaphors? Several
metaphor scholars argue that mixed metaphor is a phenomenon that CMT
cannot handle. Their argument is that, given the claims of CMT, mixed
metaphors should not occur at all (Shen and Balaban 1999). This is because
once a conceptual metaphor is activated in discourse by means of a linguistic
metaphor, that conceptual metaphor should lead to and support the use of
6.2 Mixing Metaphors 123
further linguistic examples of the same conceptual metaphor. However, in real

discourse, the argument goes, most metaphors are mixed, which indicates that
conceptual metaphors are not activated and thus do not lead to further consist-
ent linguistic metaphors of the same conceptual metaphor. In this section,
I will argue that the idea of the production of consistent and homogeneous
linguistic examples does not necessarily follow from conceptual metaphor
theory and that, in fact, the opposite is the case: given CMT, we should expect
the use of mixed metaphors in natural discourse. (This section is based on my
chapter in Gibbs, 2016.)
Kimmel’s Approach to Mixed Metaphors

Most of the metaphors used in natural discourse are of the mixed type. As
I pointed out in the present chapter and Chapter 4, this is happening at the
mental spaces level. When examining the metaphors found in natural dis-
course, we find that the metaphors in each other’s (near-)vicinity in discourse
derive from very different source domains. But we also find that this does not
really present any difficulty for us to understand the metaphors used. A version
of the CMT view is provided by Kimmel (2010) to explain this curious fact
about mixed metaphors. Kimmel argues that mixed metaphors are easily
understood in discourse largely due to various properties of discourse.
Let us now return to the specific question of why mixed metaphors are cognitively
successful. In my view, many mixed metaphors are processed unproblematically
because the grammatical structure of the passage exercises little pressure to integrate
them conceptually, and simply lets us interpret them as referring to different ontological
levels. I claim that what influences metaphor cluster processing most directly are the
relations of the clause units in which they occur. We can answer major parts of the
mixed metaphor riddle by paying attention to how adjacent metaphors are distributed
over clauses and what degree of grammatical integration these clauses show. Three
major possibilities can be distinguished. In some cases metaphors co-specify each other
within a single clause; sometimes they occur in rather tightly connected clauses, and
sometimes they only belong to a larger rhetoric structure spanning more loosely
connected clauses. (Kimmel 2010: 110)
Kimmel demonstrates these points with the example below:
Tony Blair’s criticism of EU regulations [. . .] would be laughable if it were not so two-
faced (A). While preaching the pro-business gospel (B), he has done nothing to stop
the tide of EU rules (C) and red tape (D) from choking Britain (E). (words in
boldface in the original)
Kimmel suggests that the issue and problem of mixing metaphors, that is,
conceptually integrating them, only comes up when the different metaphors
occur within a single clause, as in the case of “to stop the tide of EU rules (C)
and red tape (D) from choking Britain (E).” Whereas in the case of the
relationship between (A) and (BCDE) and (B) and (CDE) there is no pressure
to resolve the incompatibility in imagery because of the grammatical structure
of the discourse, in the case of (CDE), which is a relationship between a single
clause, there is a pressure to resolve imagistic incompatibility between “tide,”
“red tape,” and “choking Britain.” Given such examples, Kimmel draws the
conclusion: “Only the close syntactic integration of two metaphors within a
clause can enforce or foster a close integration of their semantic content qua
imagery. Where mixed metaphors occur across clauses no ontological clashes
will be felt to begin with and secondary mechanisms to keep the clash at bay
are dispensable. If this is correct, making sense of mixed metaphors is a natural
byproduct of default clause processing” (Kimmel 2010: 110).
This provides a potential explanation of why most mixed metaphors are not
taken to be cases of incompatible images and are, thus, not problematic for
metaphor interpretation.
However, convincing as far as it goes, Kimmel’s account does not explain a
number of issues relating to mixed metaphors. These include:
(1) Why are imagistically incompatible metaphors selected at a particular
point in discourse?
(2) Why are mixed metaphors so common?
(3) Why do we have cases of metaphorically entirely homogeneous
discourse?
(4) Why do we often have widely divergent source domains inserted into the
discourse?
(5) How do we comprehend mixed metaphors?
Let us briefly discuss each of the questions above.
Why Are Imagistically Incongruent Metaphors Selected

at a Particular Point in Discourse?
Given a target domain, or frame, that was introduced into the discourse, certain
elements of the domain need to be employed and linguistically expressed.
These elements are the meanings that a speaker wants to express in the course
of producing the discourse. When expressing a target domain meaning, the
speaker needs to employ an element that either comes from the target domain
directly or from a source domain that is systematically linked to that target by
means of a set of mappings. In the former case, the speaker speaks literally
(i.e., directly), in the latter, he speaks metaphorically (i.e., indirectly).
The target domain meanings form a part of aspects of the target domain,
such as progress, functioning, control, stability of structure, and so on. There
are, in many cases, conventional source domains whose main function is to
metaphorically express such aspects of target domain concepts. In other words,

they are source concepts that profile these aspects of the target domain. For
instance, the source domain of journey typically profiles the notion of
“progress,” that of machine profiles “functioning,” that of war and fight-
ing profile “control,” that of building profiles “stability of structure,” that of
the human body profiles “(hierarchical) structure” and “appropriate condi-
tion (of some structure),” and so on. In some of my publications I refer to these
as the “meaning foci” of a source domain (see, e.g., Kövecses 2000a, 2002/
2010, 2005, Chapters 1 and 4). Given these meaning foci, we get generic
metaphors like progress is (forward) motion in a journey, func-
tioning is the working of a machine, control is war/fighting,
stability of structure is the strength of a building, and several
others. These are metaphors that apply to a large number of targets; that is, they
have a wide scope (Kövecses 2000a, 2002/2010).
If this is the case, we should not expect discourses about a particular target
domain topic to be expressed by large sets of homogeneous metaphorical
expressions (i.e., expressions that belong to the same source domain). In
contrast, given a particular target domain and its various aspects, we should
expect metaphorical linguistic expressions to occur in the discourse that
capture and are based on the typical source domains that are conventionally
employed to express and capture those aspects.
To use Kimmel’s example, the writer of the passage above needs to talk
about “unnecessary paperwork, which is disliked by most people” at a certain
point of the discourse. He could use the phrase in quotation marks or some
other phrase (like “irritatingly redundant bureaucracy”), but, instead, he
chooses a metaphorical expression, red tape, which renders this meaning
appropriately in a conventionalized manner. (Actually, it is debatable whether
red tape is a metaphor. It could be argued it is a metonymy.) As a matter of
fact, the application of this phrase is just as natural as that of tide in the same
clause in the sense of “a huge amount of something.” In other words, no matter
how incompatible the images evoked by tide and red tape are, they are selected
because they can render (indirectly) the elements required by the different
aspects of the target domain at a certain point in the discourse.
Why Are Mixed Metaphors so Common?

Kimmel (2010) also observes that in the newspaper coverage of the 2004/2005
European Union referenda most of the metaphors were of the mixed type (76
percent). Mixed metaphors are common even within the same clause (see
Kimmel 2010) because the target domains, or frames, we are developing in
the course of producing and understanding (metaphorical) discourse have
many different aspects to them, and these aspects normally require different
source domains, or frames, for their conceptualization, as was noted above.
Assume, for example, that the topic of discussion is the family. The notion
(domain or frame) of family comes with a large number of different aspects
(constituted by conceptual elements), including parents, children, the raising of
children, the cost of raising children, the family as a cohesive group of people,
and a host of others. Below is a sentence I found about discussing some of
these aspects, as cited by David Brooks in the International Herald Tribune
(November 17–18, 2012): “The surest way people bind themselves is through
the family.” (my bolding, ZK) What does way have to do with bind? The word
way belongs to the motion frame, while bind to the physical joining
frame (inside the frame a group of physically connected entities).
These are incongruent images (i.e., metaphorical source domains) that create
mixed metaphors.
To account for why these and many other incongruent images are so
common, I propose, in line with the argument in the previous subsection, that
all target domain concepts that constitute the topic (i.e., subject matter) of
discourse consist of a number of different aspects, and that each of these
aspects can be conceptualized metaphorically by means of different source
domains – domains that may yield incongruent images when juxtaposed. To
take the example above, as the discussion of the family domain is unfolding,
various aspects of it are mentioned; for example, the issue of what kind of
social unity is provided by the family and the issue of how a certain state can
be achieved (“The surest way people bind themselves is through the family.”).
As regards the second metaphor of the sentence (bind), it is based on the
conceptual metaphor cohesive social groups are groups of physic-
ally connected entities. Given this metaphor, one of its mappings “the
physical joining of entities ➔ the formation of social groups” accounts for the
use of bind. Social groups constitute more or less stable states of intercon-
nected entities. The issue of how more stable states are achieved is another
aspect of the family/social group domain. And since the issue of how certain
states are achieved in general is metaphorically conceptualized as manner
(of action) is path (of motion), which is a submetaphor of (deliber-
ate) change of state is (forced) motion, we can account for the use
of the word way in the sentence above. The resulting images are fairly
incongruent when juxtaposed (way – path / motion vs. bind – the phys-
ical joining of entities / a group of physically connected
entities), but their incongruence is natural since we use different source
domains to conceptualize different aspects of the target domain, as predicted
by the notion of meaning focus. As pointed out by Kimmel (2010), this is very
common in natural discourse; it seems to be the dominant pattern – the rule,
rather than the exception.
Why Do We Have Cases of Metaphorically Entirely

Homogeneous Discourse?
If the account of metaphor in discourse as presented above is valid, it is highly
unlikely that we find metaphorically entirely homogeneous discourses. If it is
the case that different target domains have different aspects and the different
aspects are conceptualized by means of different source domains, discourse
about a particular target domain will be characterized by a number of different
conceptual metaphors, as we have seen above. However, it is clear that there
are such cases of metaphorically homogenous discourse, though their occur-
rence is not very common. Let us consider one example of such a text:
Performance targets are identical to the puissance at the Horse of the Year Show. You
know the one – the high-jump competition, where the poor, dumb horse is brought into
the ring, asked to clear a massive red wall, and as a reward for its heroic effort is
promptly brought back and asked to do it all over again, only higher.
I’ve never felt anything but admiration for those puissance horses which, not so dumb at
all, swiftly realize that the game is a bogey. Why on earth should they bother straining
heart, sinew and bone to leap higher than their own heads, only to be required to jump
even higher? And then possibly higher still.
Hard work and willingness, ponders the clever horse as he chomps in the stable that
night, clearly bring only punishment. And so next time he’s asked to canter up to the big
red wall, he plants his front feet in the ground and shakes his head. And says, what do
you take me for – an idiot? (Melanie Reid, The Times, Monday, February 4, 2008)
In this case, the same conceptual metaphor, life is a horse show, lends
coherence to a single discourse. The novel conceptual metaphor that structures
the text is characterized by a number of mappings: puissance horses are compared
to people, riders to managers, the red walls as obstacles to the targets people have
to achieve, having to jump over the obstacles to being subject to assessment,
clearing the obstacles to achieving the targets, raising the obstacles to giving
more difficult targets, the horse show to life, and so on and so forth. This
elaborate metaphorical analogy provides a great deal of structure for the text.
However, it may be even more important than the coherent organization of
the discourse that the author wishes to make a point emphatically and vividly
by the use of the homogeneous metaphor. This expressive function of the
discourse can often provide an explanation for the use of a single coherent
metaphorical image in discourse.
In addition to its organizational and expressive functions, metaphorically
homogeneous discourses can also be used for didactic purposes. Parables are
commonly used this way. Consider the parable of the sower in Matthew 13
(Biblegateway, www.biblegateway.com/passage/?search=Matthew+13&ver
sion=NIV):
1
That same day Jesus went out of the house and sat by the lake. 2 Such large crowds
gathered around him that he got into a boat and sat in it, while all the people stood on
the shore. 3 Then he told them many things in parables, saying: “A farmer went out to
sow his seed. 4 As he was scattering the seed, some fell along the path, and the birds
came and ate it up. 5 Some fell on rocky places, where it did not have much soil. It
sprang up quickly, because the soil was shallow. 6 But when the sun came up, the plants
were scorched, and they withered because they had no root. 7 Other seed fell among
thorns, which grew up and choked the plants. 8 Still other seed fell on good soil, where
it produced a crop – a hundred, sixty or thirty times what was sown. 9 Whoever has ears,
let them hear.”
Here Jesus teaches people about the kingdom of God. In doing so, he uses the
conceptual metaphor god’s words are seeds. By using a consistent image,
he feels he can convey his message more clearly to people.
Characteristic of the discourses structured by the life is a horse show
and god’s words are seeds metaphors is their conceptual homogeneity.
This is provided for by consistently applying images (elements: ideas, mean-
ings) from the same source to structure the various aspects of the target. In
such cases, as a result of the consistent structuring (based on a single source
image), no or very few mixed metaphors are found in discourse. What
discourses such as the above show is that certain discourse functions (vividly
making a point, teaching people) may provide fairly strong motivation for the
use of a single consistent metaphorical image (unlike the typical case of
metaphorically varied imagery, where the discourse has less marked rhetorical
or pragmatic functions).
Why Are Often Widely Divergent Source Domains Inserted

Into Discourse?
While in the case of metaphorically homogeneous discourse we experience a
certain degree of surprise as a result of the unexpectedly homogeneous char-
acter of the discourse, the opposite can also occur: In certain other kinds of
discourse, we can be surprised by some of the widely divergent metaphors that
are used in connection with a target domain. This element of surprise comes,
I suggest, from the fact that, first, instead of the expected (conventionalized)
metaphors associated with a target (see the “family” example above) and,
second, deliberately homogeneous metaphors chosen for a target (see “puis-
sance” example above), entirely divergent metaphors are used for the target
domain. The question is where such unexpectedly divergent metaphors come
from. (It should be noticed that here the question word “why” means some-
thing different than in the previous subsection titles; instead of the meanings
“what causes the use of mixed metaphors?” and “for what purpose do we use
mixed metaphors?,” here it means “what allows us to use mixed metaphors?”)
I argued in some recent publications (Kövecses 2005, 2010b, 2015a, and

Chapter 5) that the choice of metaphors in particular discourses is often influ-
enced by the local and global context. As we saw, the local context involves,
among other things, the immediate cultural, social, physical, linguistic, etc.,
contexts of the discourse. Let us take an example that I first analyzed in
previous publications (Kövecses 2010a,b). In it, Bill Whalen, a professor of
political science in Stanford and an advisor to Arnold Schwarzenegger in his
campaign to become governor of California, says the following about
Schwarzenegger:
“Arnold Schwarzenegger is not the second Jesse Ventura or the second Ronald Reagan,
but the first Arnold Schwarzenegger,” said Bill Whalen, a Hoover Institution scholar
who worked with Schwarzenegger on his successful ballot initiative last year and
supports the actor’s campaign for governor.
“He’s a unique commodity – unless there happens to be a whole sea of immigrant body
builders who are coming here to run for office. This is ‘Rise of the Machine,’ not
‘Attack of the Clones.’” (San Francisco Chronicle, A16, August 17, 2003)
Focusing on the second paragraph, we may note that the target domain that is
discussed by Whalen is Schwarzenegger’s uniqueness as an individual. The
statement “He’s a unique commodity” clearly indicates this. The linguistic
metaphor unique commodity is based on the people are commodities
conceptual metaphor. Next, we find the highly conventional metaphorical
expression “a whole sea of immigrant body builders,” where sea indicates
any large quantity, in this case body builders. This is an instance of the
conventional conceptual metaphor a large number of people is a
large mass of substance. Equally conventional is the linguistic metaphor
“run for office,” an instance of the presidential election is a race
conceptual metaphor. The commodity and large mass of substance
metaphors emphasize Schwarzenegger’s unique character.
Against the background of this “sea” of highly conventional metaphorical
expressions and conceptual metaphors, two nonliteral phrases at the end of the
paragraph, “This is ‘Rise of the Machine,’ not ‘Attack of the Clones,’” stand
out as unconventional and novel. In all probability, Whalen produces them,
and we understand them, because Schwarzenegger played in the first of these
movies, and we share this knowledge with the conceptualizer (Whalen) about
the topic of the discourse (Schwarzenegger). Moreover, these are movies that
everyone knew about in California and the United States in 2003; that is, they
were part and parcel of the immediate cultural context at the time.
The movie title Rise of the Machine is a widely divergent metaphor, in that it
does not come from the available set of conventional conceptual metaphors to
capture an aspect of the target domain. It derives from and is sanctioned by the
immediate context of the discourse in which the metaphor is created. Such
metaphors often provide images that may not fit the conventionally available
source domains of a given target; nevertheless, they are easily understood
through the supportive role of context. Through the notion of contextual
priming that was discussed in Chapter 5 we can account for the use of widely
divergent metaphors used in some instances of natural discourse.
How Do We Comprehend Mixed Metaphors?

The issue of mixed metaphors is largely an issue in metaphor processing. We
can think of metaphor comprehension in two ways: We can suggest that
metaphorical expressions in discourse (at the mental space level) activate
higher-level conceptual structures (frames, domains or image schemas) repre-
senting source domains that are linked to target domains. This is the more or
less accepted view among those who embrace conceptual metaphor theory (As
we saw in Chapter 4 and throughout, the three higher source domain structures
are linked in a schematicity hierarchy.)
Another view of the comprehension of metaphorical expressions in dis-
course maintains that there is no activation of source domains in most cases.
Instead, the metaphorical expressions used in discourse are associated with a
conventionalized metaphorical sense in the target domain, and it is this target
meaning that gets activated in the course of processing metaphorical discourse.
The two views have different implications for the study of mixed metaphors.
On the CMT view, whenever we come across mixed metaphors, we have to
somehow resolve the incompatibility in imagery between two source domains
that get juxtaposed. On the non-CMT view, such incompatibilities do not arise
because linguistic metaphors are processed directly – namely, in terms of their
target domain meanings. It follows that if we have no difficulties understand-
ing discourse involving mixed metaphors, it is because either we eliminate the
incompatibilities in the course of processing (for the CMT view) or because we
apply “shallow processing” (Gibbs 1999a,b), that is, we comprehend target
meanings directly (for the non-CMT view).
Given the two ways of metaphor comprehension briefly sketched above, it is
the non-CMT view that appears to be the more adequate way to account for the
processing of mixed metaphors: If people can process the majority of mixed
metaphors effortlessly, this is because they do not run into any incompatibil-
ities as a result of shallow processing, where they comprehend the metaphors
of discourse in a direct manner.
Can we nonetheless save the CMT view of metaphor comprehension for,
or in the case of, mixed metaphors? I think this is possible. My suggestion
would be that the various source domains (syntactically close to each other)
that are associated with various aspects of target domains get activated to
varying degrees in the course of metaphor comprehension. In the case of
comprehending mixed metaphors (the majority of metaphor examples in

natural discourse), the various (near-)adjacent source domains are activated
to a low degree. This way, the low level of activation for a given source does
not interfere with the low level of activation of another source. Consequently,
if the activation level of adjacent sources is low, it does not produce incompat-
ibilities or conceptual clashes between the sources. This assumes a threshold
above which source activation can produce interference or incompatibility
between two (near-)adjacent source domains. The observation that there is
no such felt incompatibility between the sources indicates that the level of
activation stays below the threshold level in most cases of mixed metaphors in
natural discourse. At the same time, for this model of comprehension to work,
the low level of source activation should be sufficient to activate source-to-
target mappings on which target meanings are based. It is this mechanism that
can function as the substitute for the “direct target meaning” of the shallow
processing view.
In other words, in this account I rely on and assume the validity of two
propositions. One is that in the case of metaphor the various source domains
and the metaphorical linguistic expressions that make them manifest are
characterized by differing degrees of conventionality. Second, the various
degrees of conventionalization for the source domains and source expressions
are correlated with differing degrees of neural activation. Lower levels of
conventionality and, hence, activation do not produce consciously recognized
incompatibility between source domains and their metaphorical expressions,
while higher levels do. In my view, this is what we find in most cases of the use
of mixed metaphors, as discussed above. This is a commonsense assumption
that I offer merely as a hypothesis for how we comprehend mixed metaphors in
discourse. The hypothesis would of course have to be tested empirically.
Consider now a case where the level of activation, in my judgment at least,
does go beyond the threshold. Take the following examples of mixed meta-
phors from one of the many websites devoted to the topic (http://therussler
.tripod.com/dtps/mixed_metaphors.html):
If they do that, they might as well take the open door policy and throw
it right out the window!
The example is about policy making. This involves the kinds of policy one
pursues and whether the policies used are successful or appropriate or not. The
phrase open door policy is based on the source domain of container
(a government is a container) and seeing (knowing is seeing).
The phrase throw it out of the window implies the source domain of dis-
carded objects (in the metaphor unnecessary ideas are discarded
objects. How do we know that the activation level goes beyond the thresh-
old? We know it because although we understand the target domain meaning
of the sentence easily, the sentence also produces a humorous effect. As is

well-known, humor often involves some incongruence between two images
(such as source and target or, in the case of mixed metaphors, two adjacent
source domains). The humorous effect can only be produced by a very high
level of activation for the sources as well (and not just the target). Without the
high level of activation of the two sources, we would easily gloss over the
obvious incompatibilities. In cases like this, the humorous effect produced by
many mixed metaphors can help us determine the level of activation of the two
adjacent source domains.
Some opponents of (parts of ) conceptual metaphor theory (e.g., Shen and
Balaban 1999) hold that it is only the widespread application of metaphorically
homogeneous discourse that could prove source domain activation in the use
of conceptual metaphors. Since, however, such metaphorically homogeneous
discourses are rare, they suggest that one of the major claims of CMT (i.e.,
simultaneous activation of source and target) is called into question.
Why are metaphorically homogeneous discourses so rare? In all probability,
this is because, in line with the proposal offered above (in the first two
subsections of this section), most metaphorical source domains profile a
particular aspect (or aspects) of that source domain. But if they do occur, this
indicates that a single source domain is used to characterize an entire target (or
a large part of it). In this case, we find the full (or rather, nearly full) activation
of that source for the target in question (cf. the horse show example). In
terms of the activation model I am suggesting, this would mean that the
activation level of the source is above the threshold and remains high through-
out the relevant parts of the discourse. This can be achieved under two
conditions.
One is that the source domain is conceptually complex enough to be
mapped onto a several different aspects of the target. For instance, the
conceptual complexity of the horse show source domain seems to be
sufficient to present several rather different aspects of life that could also
be rendered by means of several different source domains with the appropriate
meaning focus.
The other condition is that for some communicative or stylistic reason it is
advantageous for the speaker/conceptualizer to conceptualize (and talk about)
the target via the same source domain. Making a point vividly and forcefully,
as in the case of the horse-show example, and accomplishing a didactic
purpose, as we saw in the case of Jesus’ parable of the sower, may be one
such reason. This second condition is clearly needed: Most cases of metaphor-
ically homogeneous discourse have an added communicative or stylistic com-
ponent (such as convincing, persuasion, irony, humor, esthetic effect) in a way
that is not present in discourses that fit the dominant pattern of metaphorically
non-homogeneous discourse.
6.3 Conceptual Metaphor and Conceptual Integration 133
Mixed Metaphors in Conceptual Metaphor Theory

In the section, I argued that CMT can handle mixed metaphors in a natural
way. Its naturalness arises from the fact that many common target domains
have a number of different aspects to them that can be, and often are, concep-
tualized in terms of a variety of potentially very different source domains that
we use at the mental space level in naturally occurring discourse. The use of
these different source domains leads naturally to the use of mixed metaphors.
In a way, most communicatively and stylistically neutral cases of discourse can
only be metaphorically non-homogeneous discourses that consist of a variety
of source images, that is, mixed metaphors.
We do not have any difficulty processing such discourses. My suggestion
here was that this is possible because, in most cases, source domains are
activated at a low-level. This low level of activation for a particular domain
does not interfere with the activation of an adjacent source domain in the
discourse. However, the low level of activation should be sufficient enough to
activate source-to-target mappings that can result in the required meanings
needed to understand and produce discourse. All of these mental activities are
happening in working memory – the site for mental spaces.
Clearly, this model of comprehending mixed metaphors is just a hypothesis.
The model involving varying degrees of activation of source domains seems to
provide a reasonable framework for the comprehension of mixed metaphors
and at the same time to be compatible with conceptual metaphor theory at large
and with its present extension.
6.3 Conceptual Metaphor and Conceptual Integration

What is the relationship between conceptual metaphor and conceptual integra-
tion? A clear answer to this question is given from the perspective of concep-
tual integration theorists in the following passage: “Contemporary accounts of
metaphor and analogy have focused on structure-mapping from a source (or
base) onto a target. Such mappings can exploit existing common schematic
structure between domains, or project new structure from the source onto the
target. The work on conceptual blending has shown that in addition to such
mappings, there are dynamic integration processes which build up new
“blended” mental spaces. Such spaces develop emergent structure which is
elaborated in the on-line construction of meaning and serves as an important
locus of cognitive activity” (Turner and Fauconnier 2000: 133).
On this view, the idea of metaphor as structure mapping from source to
target in CMT should be supplemented by blended mental spaces that are the
result of dynamic, online integration processes as described by conceptual
integration theory (CIT). The two-domain model of CMT (as described by
Kövecses 2002/2010; Lakoff 1993; Lakoff and Johnson 1980, 1999) should
give way to the “network model” that comprises several input spaces, a generic
space, and the blended space.
More recently, Fauconnier and Turner (2008: 53) took this argument one
step further as we can see in the passage below: “What we have come to call
“conceptual metaphors,” like time is money or time is space, turn out to
be mental constructions involving many spaces and many mappings in elabor-
ate integration networks constructed by means of overarching general prin-
ciples. These integration networks are far richer than the bundles of pairwise
bindings considered in recent theories of metaphor.”
In other words, the suggestion is that what are known as the typical
examples of conceptual metaphors are not simply “pairwise bindings” but
are the products of “elaborate integration networks.” All conceptual metaphors
are in fact blends.
Most generally and significantly, CIT researchers claim that metaphors,
together with analogy, framing, metonymy, etc., are surface cognitive phe-
nomena that “are consequences of the same basic human ability for double-
scope blending” (Fauconnier and Turner 2008: 54). In other words, metaphors
are relatively surface products of the deep, underlying process of conceptual
integration.
I agree with many of these suggestions mentioned above. For example,
I agree that in many cases CMT should move beyond structure-mapping and
inference transfer. I also accept the criticism that in many cases mappings
occur not only in a source-to-target direction. And most obviously, I also think
that blended spaces are clearly needed to account for a variety of examples that
CMT alone could not account for.
Nonetheless, in this section, I offer a somewhat different view on the
relationship between CMT and CIT than what we saw above. I group what
I have to say about the relationship around three topics. In particular, the first
of these concerns the role and importance of structure mapping both in CMT
and CIT. The second has to do with the claim that blending is somehow a more
basic, deeper cognitive operation than metaphor. The third is the issue whether
blending is indeed a more complex cognitive phenomenon than metaphor. My
answer to each of these questions will be a “partly yes, partly no” response.
Conceptual Metaphors and Generic Space

When CIT deals with cases where there is a source domain (input space 1) and
a target domain (input space 2) in “standard” CMT, it operates with the notion
of similarity. The similarity is spelled out in the generic space. The generic
space consists of information that the source and the target share. For example,
the sentence “This surgeon is a butcher” assumes the conceptual presence of
GENERIC SPACE
A person
Employing a sharp tool to a body for a
purpose
SOURCE DOMAIN: BUTCHERY TARGET DOMAIN: SURGERY
The butcher The surgeon

The tool used: the cleaver The tool used: the scalpel
The animal (carcass) The human being
The commodity The patient
The abattoir The operating room
The goal of severing meat The goal of healing
The means of butchery The means of surgery
The butcher/The surgeon

The tool used: the scalpel
The human being
The patient
The operating room
The goal of healing
The means of butchery

Incompetence
BLEND
Figure 6.1 Generic space and similarity in “This surgeon is a butcher.”
some underlying similarities given in the generic space: In both inputs, there is
a person who employs a sharp tool to a body for some purpose. This is
represented in Figure 6.1 above, where the entire CIT network is given, but
for now we only focus on the notion of the schematic knowledge shared by the
input spaces.
In other words, CIT spells out the similarities between source and target at a
schematic level. The similarities form the basis for bringing together the two
domains in the network. According to CIT scholars, this allows conceptuali-
zers/speakers to set up correspondences, or mappings, between elements of the
domains. That is, given the generic space above, we can map the butcher onto
surgeon, the cleaver onto the scalpel, etc.
A major consequence of this situation, however, is that since it is the generic

space that makes it possible for the inputs to appear in the network, it seems to
follow that CIT can only apply to resemblance metaphors. (For a classification
of metaphors in the CMT approach, see Grady 1999.) This is because the other
major type of metaphors identified by CMT, correlation metaphors, are not
based on any similarity – they are based on correlations in experience. It is
unclear how CIT can accommodate correlation-based metaphors as described
in CMT. If it cannot, it appears to weaken the claim made by CIT scholars that
all conceptual metaphors can be analyzed as blends. Actually, this issue was
also noticed by Grady (2005), who proposed that correlation metaphors should
be regarded as input spaces in integration networks. This would mean that
correlation-based metaphors would not be consequences of double-scope
blending but would serve as their essential ingredients.
Abstract Concepts in the Network

As we saw, blending assumes at least two input spaces (which can be related as
source and target), a generic space (identifying the similarities), and the blend.
If we insist on the initial and simultaneous presence of all four of these in the
network, and especially the second input (the target) in cases of metaphorical
conceptualization, we encounter a difficulty in accounting for the initial
emergence of abstract concepts. How do abstract concepts emerge in the CIT
framework? If the second input, corresponding to the abstract target domain, is
already there in the network, we are assuming its presence from the very
beginning – instead of explaining its emergence. Notice that the problem
cannot be avoided by suggesting that the abstract input derives from a prior
network, since that network must also contain a second input that is already
abstract. But how did it come into being to begin with?
As a matter of fact, this is also a problem for CMT. CMT with its two-
domain model (A IS B) suggests that the source is typically a physical concept
and the target an abstract one. We can ask the same question as above: But
how did the abstract one initially emerge? The A IS B formula is misleading
because it assumes that the A, the abstract concept, is already there. After it
emerges, we can legitimately propose that new aspects of the abstract target are
created by means of new source domains. But how do they arise to begin with?
The challenge then, both for CIT and CMT, is to offer an account of how
abstract target concepts emerge from scratch.
Let us consider the possibilities that are commonly offered for the creation
of conceptual metaphors with an A IS B structure, where A is abstract and B is
concrete. One of the possibilities is resemblance. In resemblance metaphors,
the metaphor is based on some similarity between A and B. The similarities
can be of various kinds (see, e.g., Kövecses, 2002/2010; Lakoff and Johnson,
1980). Strictly speaking, resemblance is not a truly viable option to explain the
problem, since the two concepts must already exist before any similarity is
perceived between them (see above).
Correlation in experience, however, is more promising. In primary metaphors
such as purposes are destinations, the destination can be generalized
into a purpose concept: the desire to get to a physical location can lead to the
emergence of the notion of purpose that is not limited to getting to a location
but applies generally (on this process, see Kövecses 2013 and Chapter 3.)
Another, and less recognized, way for abstract concepts to emerge is
through schematization. In several publications (see Kövecses 2011c,
2015c), I argue that metaphors such as heaven is an ideal physical
place, causation is progeneration, and providence is nurtur-
ance are based on the source domain schematized into the target. This is
happening via a metonymic process. The specific rich imagery of the source is
bleached out to yield such abstract (target) domains as heaven, causation,
and providence.
When Do Metaphoric Blends Arise?

According to CIT, all conceptual metaphors are blends. Let us examine this
issue by first looking at an example. Take the emotion is an opponent
metaphor, as exemplified by sentences such as “He was struggling with his
emotions” and “She was overpowered by emotion.” There are two frames
here: the opponent frame and the emotion frame. In the opponent frame,
we have two participants, opponent 1 and opponent 2, who are struggling with
each other. They both want to defeat each other. One opponent wins, the other
loses. In the emotion frame, we have the self and the emotion, and the self is
trying to control his/her emotion. Either the self or the emotion will control the
other. Thus we have the mappings:
opponent 1 ➔ self
opponent 2 ➔ emotion
struggle ➔ attempt at control
winning ➔ gaining control
losing ➔ losing control
These elements in the two frames are compatible with each other, and so the
elements in the opponent frame map straightforwardly onto the elements in
the emotion frame.
In a CIT approach, it could be suggested that the two input spaces form part
of a “double-scope” network, in which the matching elements in the two
spaces, or frames, would be conceptually integrated in the blended space as
follows:
opponent 1 / self
opponent 2 / emotion
struggle / attempt at control
etc.
However, I do not think that this would yield a substantially different result
from the suggestion made by CMT. These are cases where two frames have the
same schematic structural relations between the elements, and so the elements
in the frames are compatible with each other in a straightforward manner. In
the CIT account, the mappings from source-to-target are simply renamed as
blended counterparts in the blend. It is not clear whether the idea of “maps
onto” and “are blended/integrated/fused” mean something very different in
such cases. It can thus be suggested that the two accounts are simply termino-
logical variants.
True cases of blending arise when, over and above the compatibilities, there
is some incompatibility between the two frames. Consider the well-known
example offered by Turner and Fauconnier (2000) and briefly commented on
above in the chapter, God, he was so mad smoke was coming out of his ears.
The sentence is based on the anger is heat conceptual metaphor (in its
two versions: hot fluid in a container or fire). There are several
mappings between the angry person and heat (either boiling pot or fire) that
obtain between the compatible elements in the two domains:
the container ➔ the angry person’s body
the heat ➔ the anger
the degree of the heat (of fluid) ➔ the intensity (of anger)
the cause of the heat ➔ the cause of anger
etc.
These elements in both frames constitute a structured whole. The two struc-
tures are compatible with each other, just like in the emotion is an oppon-
ent metaphor we saw above. However, in the “smoke was coming out of his
ears” example, there is a major incompatibility. Here, incompatible elements in
the source and target are brought together to form a conceptual unit that is not
part of either the source or the target: The element that indicates a high degree
of heat (the smoke or fume or steam) is conceptually integrated with an
element of the target – the angry person’s body (more specifically, the head).
The result is a smoking, fuming, or steaming head (of the angry person). The
sentence creates a new frame (a blended space) in which smoke/steam is
coming out of the person’s ears. The new frame results from merging two
otherwise incompatible elements from the source and target frames.
But there are even more complicated cases that we can account for by
making use of the idea of blending as arising from some incompatibility
between the elements of the source and target. Take the surgery is butch-
ery metaphorical blend discussed previously. We noted above that much of
the conceptualization involved in this example is very much like a straightfor-
ward set of metaphorical mappings: certain elements of the source correspond
to certain elements of the target. Here, we have a systematically organized
frame with elements, butchery, mapping onto the elements of another
systematically organized frame, surgery.
As Figure 6.1 above shows, on the blending view, in addition to the two
input spaces of butchery and surgery that are connected by a set of
mappings, or correspondences, as above (except for the last correspondence),
we have a generic space in which there is a person who employs a sharp tool to
a body for a purpose. There is also a blended space. This space inherits from
the source input the butcher and the means of butchery and from the target
input the surgeon, the patient, some tool, the operating room, and the goal of
healing. Thus, in the blend there is a surgeon in the role of a butcher who uses a
tool and the means of butchery for the purpose of healing a patient. But, of
course, the surgeon who uses the means of butchery cannot do a good job in
trying to heal a human patient. The blend set up this way leads to the
interpretation of the surgeon as being ineffective, nonprofessional, sloppy,
careless, etc., and, ultimately, incompetent.
Kövecses (2011b) uses the idea of “meaning focus” to analyze the example
“This surgeon is a butcher.” By employing the notion of meaning focus (see
Kövecses 2000a, 2002/2010), we could eliminate the problem associated with
“standard” conceptual metaphor theory: the problem that, on that analysis,
there is no account of why the feature of sloppiness or carelessness (or
incompetence) is mapped onto the surgeon. The view based on the meaning
focus of the source domain would maintain that the feature is mapped onto the
target because it is one of the meaning foci associated with butchers, as
indicated by the various senses of the word as defined by Merriam-Webster:
1 a: a person who slaughters animals or dresses their flesh b: a dealer in meat
2: one that kills ruthlessly or brutally
3: one that bungles or botches
4: a vendor especially on trains or in theaters
Sense 3 clearly indicates that butchers are regarded as inherently incompetent.
Given this conventionalized sense of the word and given that source domains
map their main meaning focus (whose selection from several potential foci
may depend on the context) on the target, we can understand why the meta-
phorical statement means what it does.
In addition, to account for the meaning of the sentence, we also need some
metonymies. There is a metonymic relationship between the category as a
whole and the property as a part, yielding the metonymy category for
its property (see Kövecses and Radden 1998). That is, the word butcher is
used in the sentence to metonymically indicate incompetence.
Let us return to the issue of the butcher being “careless, sloppy, imprecise,”
that is, incompetent. Butchers are not inherently incompetent, but sometimes
we take them to be such. Why? The reason may be that the actions performed
by the butcher appear to be sloppy in contrast to the surgeon. We compare the
butcher’s actions with the “precise” and “refined” actions of the surgeon (cf.
the phrase “with surgical precision”). In other words, we interpret how the
butcher works in reference to the surgeon’s work, that is, with the surgery
frame in the background. Therefore, we extend the primary meaning of the
word butcher (“who slaughters animals and dresses their flesh”) to “incompe-
tent.” This newly derived meaning will then be projected to and will charac-
terize the particular surgeon as well. We can think of the projection as a case of
conceptual integration. The projection goes to a new space, the blended space,
where the “careless, sloppy work” of the butcher replaces the “precise and
refined work” of the surgeon. In this way, the blend contains mostly infor-
mation that the surgery frame contains, with the major difference that the
particular surgeon will here be regarded as doing “careless and sloppy work”
and, hence, “being incompetent.” The surgeon in the blend assumes the
contextually defined new meaning focus of the butcher. We can diagram this
as in Figure 6.2 below on the next page.
The above solution to the question of how the meaning of the sentence
arises is different from the CIT solution proposed by Grady et al. (1999). In
CIT, the essential elements of the blend were the means of butchery and the
goal of surgery, as well as the conflict between the two, leading to the property
of “incompetence.” In the present suggestion, the property of “incompetence”
is projected into the blend from the input space of butchery. The property of
“incompetence” emerges in the concept of butchery in light of and against
the background of the concept of surgery, and then it is projected into the
blend, in which the property will now characterize the surgeon. That is, critical
conceptual material in the blend will derive from the source. Thus, if the
contextually defined meaning focus of the source domain of butcher is that
he does careless, sloppy work (relative to the surgeon), it will carry over this
property to the target (and to the blend, as above).
Let us now turn to a case where the incompatibility is between two stories,
as in parables. In some parables, when a new didactic point is made by the
speaker, the two stories involved in the parable are incompatible, and despite
the several systematic correspondences between the source and the target, the
source story does not match the target story in certain respects. This may cause
surprise or difficulty in interpretation (see the section on “mixing metaphors”
above). The Parable of the Workers in the Vineyard in the New Testament of
the Bible is a good example of such a situation:
GENERIC SPACE
A person
Employing a sharp tool to

a body for a purpose
INPUT 1 – SOURCE INPUT 2 – TARGET

DOMAIN: BUTCHERY DOMAIN: SURGERY
The butcher The surgeon

Precise work
The butcher The butcher/ The surgeon

Careless, sloppy work Careless, sloppy work
NEW SOURCE DOMAIN: BUTCHERY BLEND
Figure 6.2 The new surgeon as butcher blend in the “meaning

focus” view
20 ”For the kingdom of heaven is like a landowner who went out early in the morning
to hire workers for his vineyard. 2 He agreed to pay them a denarius[a] for the day and
sent them into his vineyard.
3
”About nine in the morning he went out and saw others standing in the marketplace
doing nothing. 4 He told them, ‘You also go and work in my vineyard, and I will pay
you whatever is right.’ 5 So they went.
“He went out again about noon and about three in the afternoon and did the same thing.
6
About five in the afternoon he went out and found still others standing around. He
asked them, ‘Why have you been standing here all day long doing nothing?’
7
”‘Because no one has hired us,’ they answered.
“He said to them, ‘You also go and work in my vineyard.’
8
”When evening came, the owner of the vineyard said to his foreman, ‘Call the workers
and pay them their wages, beginning with the last ones hired and going on to the first.’
9
”The workers who were hired about five in the afternoon came and each received a
denarius. 10 So when those came who were hired first, they expected to receive more.
But each one of them also received a denarius. 11 When they received it, they began to
grumble against the landowner. 12 ’These who were hired last worked only one hour,’
they said, ‘and you have made them equal to us who have borne the burden of the work
and the heat of the day.’
13
”But he answered one of them, ‘I am not being unfair to you, friend. Didn’t you agree
to work for a denarius? 14 Take your pay and go. I want to give the one who was hired
last the same as I gave you. 15 Don’t I have the right to do what I want with my own
money? Or are you envious because I am generous?’
16
”So the last will be first, and the first will be last.” (Matthew 20:1–16, New
International Version)
Most people, both in Jesus’ time and today, are surprised, and even find it
unfair, that Jesus gives the same pay to those who worked the whole day and
those who worked only one hour. This can happen because Jesus changes the
source domain story according to the target story. The target story is about the
kingdom of heaven, where all people are invited (just like the workers by the
landowner) and receive the same “pay.” However, the pay is not financial
reward but forgiveness and mercy that all people who ask for it will get in
equal amount. In other words, the kingdom of heaven is only partially similar
to our world, and Jesus adjusts the source story according to how things work
in the kingdom of heaven, the target story.
I suggest that the cognitive process that produces such an interpretation is
conceptual integration (see, e.g., Turner 1996). In blending, two partially
similar, thus essentially compatible, domains are merged, rather than the
source simply structuring the target. In my view, this merging is possible
and arises when there is some incompatibility between the (otherwise similar)
source and target. The incompatible element in the target (i.e., the element that
does not match an element in the corresponding source) will figure importantly
in the intended interpretation of the parable. In more specific terms this means
that the speaker will use elements of the source that are not coherent with the
rest of the source but that fit the target. Technically, this is achieved by the
speaker projecting the target-compatible (but source-incompatible) elements
into a blended conceptual space, where the target-compatible elements (every-
one gets the same “amount” of forgiveness and mercy) are fused with the
source-incompatible ones (everyone gets the same wage, regardless of how
many hours they work). Thus, in the blend all workers get the same amount of
forgiveness and mercy, no matter how long they have worked for it (i.e., when
they requested it).
In sum, blending is needed when incompatible elements between two
otherwise partially compatible domains are used in meaning construction.
Metaphors and Blends in Context

That the creation of metaphors is heavily influenced by contextual factors is by
now a well-established fact (for recent statements on this issue, see Kövecses
2010b, 2015a, and Chapter 5). The interesting question is whether the process
of conceptual integration undergoes the same kind of contextual influence as
metaphor does. It is an interesting question because conceptual integration is
claimed to be a more “basic,” “elaborate,” and “deeper” “underlying” cogni-
tive operation than conceptual metaphor (see Fauconnier and Turner 2002,
2008). Is this indeed the case? More specifically, is it or is it not the case that
the “deeper” process of blending is subject to the same contextual influences as
the “superficial” phenomenon of metaphor? This is an issue I raise in my
2015 book Where Metaphors Come From, and I borrow materials from that
discussion here.
We can ask why particular networks (generic space, several input spaces,
and a blended space) are formed and expressed the way they are. Why do they
have the input spaces and the particular figurative linguistic expressions they
do? Below, I discuss some examples of blending and suggest that contextual
factors often play a role in the selection of input spaces and the linguistic
expressions that realize them.
First, let us take an example analyzed by Coulson and Oakley (2000):
“Tennessee Tramples Kentucky,” which is a newspaper headline. Coulson
and Oakley correctly point out that the blended space created by this sentence
is interpretable because of the metonymic relationship between elements in the
blend and elements in the inputs. The specific metonymy would be a state
for a university in that state and for the football team of
that university. In addition, we also have the conceptual metaphor sport
is combat as part of the network (as source and target spaces). However, this
does not explain the use of the word trample in the sentence out of the many
potentially available words (e.g., defeat, beat, overpower) that could also be
employed. But the influence of the immediate linguistic context does seem to
provide some justification or motivation for the use of the word. Tennessee
alliterates with trample. This is an extremely superficial contextual influence
involving phonological form, but it appears to be sufficient to motivate the
choice of the word trample out of a number of potential words in this
metaphorical blend.
Second, similar to the trample example above, Aitchison (1987) mentions
several newspaper headlines for one football team defeating another:
“Cowboys Corral Buffalos,” “Clemson Cooks Rice,” “Air Force Torpedoes
the Navy,” and “Cougars Drown Beavers.” The issue here is why in one case
the verb used for defeat is trample, in another it is corral, in a third it is cook, in
a fourth it is drown, and so forth. The answer again has to do with context. On
the CIT analysis, for example in the case of “Cowboys Corral Buffalos,” in the
blend we would have a university team blended with an opponent and
cowboys, another university team, blended with the other opponent and
buffalos, and, finally, the action of defeating with corralling. But, again, why
is corral used out of many other options? The reason seems to be that it is the
semantic content, or meaning, of the immediate linguistic context (and not its
phonological shape, unlike the previous example above) that facilitates the
choice of a metaphorical expression: in the Wild West frame, cowboys are in
the business of corralling animals. The same kind of explanation applies to the
other cases.
Third, Coulson and Oakley (2003) examine still another headline: “Coke
Flows Past Forecasts: Soft Drink Company Posts Gains” from USA Today.
They suggest that the literal predication of the drink Coca Cola (namely, that it
flows) and the metaphoric predication of flow for the Coca Cola Corporation’s
profit are blended in the blended space. In this case, we seem to have double
motivation for the use of flow in the network. One is of course the immediate
linguistic context with the word coke, which has the property that it flows. The
other is the general topic of the discourse, Coca Cola, which is a liquid. Thus,
here, there is both the immediate linguistic context and the general topic, or
subject matter, of the discourse (a liquid) that is at work in motivating the
particular blend that is created.
Fourth, let us take one of the best known examples of blending: “If Clinton
Were the Titanic, the Iceberg Would Sink” (Turner and Fauconnier 2000).
This is a complex example whose analysis by Turner and Fauconnier shows
the full power of CIT. What their analysis does not show in an explicit and
systematic way, however, is why we have the particular input spaces that we
do in the network of spaces. To explain this, we need a model of meaning
making that takes seriously and includes a theory of contextual influence on
the construction of meaning – both for metaphor and blending (see Kövecses
2015a). Actually, Turner and Fauconnier do mention that this particular
blend could not have arisen in 1992 without the Titanic catastrophe to begin
with and without the wide popularity of the movie inside the Washington
Beltway during the Clinton presidency. However, this somewhat anecdotal
comment does not substitute for a full theory of contextual influence on
meaning construction. Therefore, I would add to the comment made by
Turner and Fauconnier that people often produce metaphorical blends by
relying on “context models” (on this, see Van Dijk 2009) that help them
activate aspects of context (i.e., contextual factors), which results in introdu-
cing an input space (typically a source concept) into the network (see
Kövecses 2015a).
In summary, over and above providing deep conceptual analyses of blends

and conceptual metaphors, we also need to be able to account for why particular
input spaces, source domains, and linguistic expressions are used in a network or
conceptual metaphor. The brief study of the four examples illustrates that in
many cases this is possible, and that such considerations add a new and
systematic dimension to our analysis. The cognitive process of blending may
be, in some sense, more basic and deeper than metaphor, but this does not make
its products, the blends, immune to contextual influence. Blends seem to be just
as much prone as metaphors to the various contextual effects in the creation of
input spaces and the linguistic expressions based on them.
Blends in the Multilevel View of CMT

Researchers in CIT emphasize that blending is a dynamic, online process. It
takes place in real discourse in specific communicative situations. It works
with mental spaces that speakers create online. The speakers project elements
from the various spaces in the network into the blended space and also from
the blend into the various input spaces. All of this can happen due to the
plasticity of the human conceptual system and the dynamic nature of thought.
In the multilevel view of metaphor, as proposed here, the suggestion is,
essentially, that metaphor as a conceptual phenomenon can be identified at
four different levels of schematicity; in particular, at the level of IMAGE
SCHEMAS, DOMAINS, FRAMES, as well as MENTAL SPACES. The first
three give us the basic patterns of conceptual systems, as they are derived from
our experiences in the world. They are in long-term memory and they provide
the essential content elements of meaning. The IMAGE SCHEMAS,
DOMAINS, and FRAMES together form the conceptual content of conceptual
systems, and, importantly, they are decontextualized conceptual structures.
However, when speakers engage in real discourse in specific communicative
settings, they mobilize, exploit, and manipulate this large amount of tacit
knowledge. This is the conceptual stuff with which they can work (i.e., this
is what they have at their disposal). They exploit and manipulate this material
given their specific goals and constraints in the situation at hand. Speakers
perform this complex cognitive task with the help of MENTAL SPACES.
Most relevantly, of course, for the present chapter, speakers sometimes feel
that they can best achieve their communicative goals if they use particular
metaphorical blends in the course of online communication. This explains why
(both) blends (and metaphors) are subject to contextual influence, as we saw in
the previous section. The contextual influence happens, quite naturally, at the
level of MENTAL SPACES that are structured in part by the higher level
conceptual units.
Some Issues
There appears to be some rivalry between CMT and CIT. Many scholars
believe that they have to choose one or the other as their research tool because
one is somehow better than the other. I believe that the conclusions that we can
draw from the preceding analyses defuse this kind of tension: there is abso-
lutely no reason to think that CMT is better or that CIT is better. They very
much depend on and complement each other. One without the other would be
much weaker. This conclusion is shared by Grady et al. (1999: 120), who state:
“. . . we propose that because they tackle different aspects of metaphorical
conceptualization, the two frameworks are largely complementary. The con-
ventional conceptual pairings and one-way mappings studied within CMT are
inputs to and constraints on the kind of dynamic conceptual networks posited
within BT [Blending Theory].”
However, the reasons for the complementary relationship between CMT and
CIT I discuss in this section are somewhat different from theirs, in that I place
the emphasis on different aspects of the relationship within a larger, multilevel
and contextual view of metaphorical conceptualization.
Compatibility Let us begin with the issue of the role and importance
of structure mapping in CMT and CIT. Although CIT theorists urge CMT
researchers to move beyond simply studying pairwise correspondences
between two domains, it is clear that structure mapping of the kind CMT is
interested in is just as important for CIT as it is for CMT. We saw above in a
number of examples that shared structure is essential not just for metaphorical
mapping (from B to A) but also for all the other kinds of mapping and
projections in CIT that characterize metaphorical blends. The mental spaces
in double-scope networks allow for mappings not just from the source domain
to the target but also mappings from both (or more) of the input spaces into the
blended space and even from the blend into the input spaces. All of these
mappings (or projections), which are additional to the standard metaphorical
mappings (i.e., from source to target), assume compatible input spaces in the
sense that the additional mappings or projections are the result of incompat-
ibilities that themselves derive from systematic compatibilities (as captured by
systematic source-to-target mappings). This means that we cannot focus on the
incompatibilities without at the same time focusing on what is compatible
between two domains. In order to see the incompatibilities, we also have to see
the compatibilities. We cannot figure out what the incompatible elements in
blends are without identifying what the compatibilities are on which they are
built. Thinking of the relationship this way, we can suggest that CMT, on the
whole, is the study of compatibilities, whereas CIT is the study of the incompat-
ibilities that result from those compatibilities (as found in the source-to-target
mappings). Metaphorical blends are only imaginable if they assume compatible

source and target domains. Incompatibilities in themselves cannot constitute
metaphorical blends. (Centaurs are blends, but they are not metaphorical ones).
Basicness We can now turn to the second issue: whether blending is

indeed a more basic, deeper, hence more general, cognitive mechanism than
conceptual metaphor. This issue is a matter of perspective. We can approach it
from the level of mental spaces, on the one hand, or from the level of image
schemas, domains, and frames, on the other. Accepting all the claims of CIT
and approaching it from the perspective of mental spaces, it can be suggested
that the level of mental spaces and the network models within it are indeed
underlying many different conceptual structures (metaphor, analogy,
metonymy, frames, etc.) that can thus be regarded as surface cognitive phe-
nomena. But we can look at the issue differently. We can suggest that it is the
large body of such conceptual structures as IMAGE SCHEMAS, DOMAINS,
and FRAMES that constitutes the backbone of the conceptual system. And this
backbone is based on a schematic hierarchy of coherently structured human
experience – that is, the large body of compatibilities (between these struc-
tures) that the network model assumes (and of course changes, alters, and
modifies in the course of its workings). As it was noted in the previous
subsection, the incompatibilities and clashes that are resolved in the blends
are all made possible by the compatibilities (the schematic vertical similarities
between image schemas, domains, and frames). Ultimately, the question is:
What is more basic – seeing the schematic similarities or creating the dissimi-
larities? Compatibility or incompatibility? Order or chaos? Philosophers could
probably argue for both. I take it that it is the similarity- and analogy-based
conceptual structures that are fundamental to human cognition. But what
makes us truly human is that we can create the incompatibilities based on
them – and thus can easily understand them. On this I agree with CIT.
There is also some evidence to support the suggestion that, in a sense,
conceptual metaphor is more basic than conceptual integration: It is that
conceptual metaphors have been shown to be compatible with computational
modeling of neuronal connections, specifically “mapping circuits” that link
two groups of nodes (source and target) to each other (Gallese and Lakoff
2005; for compatible evidence, see also Bambini et al. 2011). The elements in
one group of nodes are connected to elements in the other group of nodes (this
is what we call “mappings”) in a systematic and predictable way. There is no
clear neurocomputational evidence supporting blending. It remains unclear
what routines and neural circuits might predict which element in one space
should be mapped onto which element in another space in the network model.
The whole point of blending is to allow the mappings to go in unpredictable
and creative ways. The processes of blending take place online, dynamically,
and in a creative fashion, and they are governed by a variety of contextual

factors, including but not limited to the contextual factors mentioned above in
the chapter.
Complexity The third issue I’d like to comment on is that of cognitive

complexity in CMT and CIT. To quote Fauconnier and Turner (2008: 53)
again: “What we have come to call ‘conceptual metaphors,’ like TIME IS
MONEY or TIME IS SPACE, turn out to be mental constructions involving
many spaces and many mappings in elaborate integration networks con-
structed by means of overarching general principles. These integration net-
works are far richer than the bundles of pairwise bindings considered in recent
theories of metaphor.”
It would be hard not to agree with this view in CIT. The analysis of specific
metaphorical blends in the CIT literature brought to light an enormous amount
of cognitive complexity in the course of metaphorical meaning construction.
Compared to this, the usual CMT analyses are simple, and sometimes even
appear simplistic. But the question is: Can we compare and contrast the
complexity of metaphor with the complexity of blends? I do not think we
can. The reason is that the cognitive complexity that characterizes metaphors is
an entirely different sort from that of blends.
The “pairwise bindings,” the mappings, or correspondences between two
structured conceptual organizations, are intended to capture the structure of
conceptual metaphors at the levels of IMAGE SCHEMAS, DOMAINS, and
FRAMES. Conceptual metaphors laid out in this format (i.e., as sets of
mappings) display the systematic and coherent compatibilities between two
IMAGE SCHEMAS, between two DOMAINS, or between two FRAMES.
These linked conceptual structures are in long-term memory, and are decon-
textualized and conventionalized. They are based on our bodily-cultural (or
cultural-bodily) experiences. And, last but not least, they are arranged in
vertical hierarchies based on schematicity. This is a huge system of concepts
that needs to be activated to function online. We are also probably predisposed
to use this system in neutral contexts. (This is what I called metaphorical
concepts in the “conceptual-cognitive context” in Kövecses 2015a and Chap-
ter 5.) Now, at the level of mental spaces, that is, in the course of online
metaphorical conceptualization in actual communicative situations, the net-
work model takes over. At this fully contextualized level, we are submerged,
and sometimes even, revel, in all the cognitive complexities that CIT is
discussing, including, importantly, the large-scale exploitation and mobiliza-
tion of the huge and dormant metaphorical system.
But the two main levels (the “aggregate” level of IMAGE SCHEMAS,
DOMAINS, and FRAMES, on the one hand, and MENTAL SPACES, on
the other) should be distinguished for theoretical purposes, since they serve
6.4 Conclusions 149
and perform very different cognitive tasks. For this reason, we cannot expect
the large dormant system to perform all the complex cognitive tasks that the
other fully specific and fully contextualized system can. I would guess that the
simplicity of the large system is by design: to make sure that we are capable of
keeping it in long-term memory for the purposes of mobilizing it for the
intricate application of the double-scope networks in working memory.
In conclusion, rivalry is unjustified and unnecessary. We are doing different
but complementary things within a unified and many-layered system of meta-
phorical conceptualization. What CIT researchers do at the level of mental
spaces would not be possible without the findings of the new, extended CMT,
and the work in CMT done at the level of image schemas, domains, and frames
would be incomplete without the contributions of CIT.
6.4 Conclusions
The phenomenon of conceptual metaphor is simultaneously offline and online.
In the course of using conceptual metaphors, offline conceptual structures in
long-term memory (image schemas, domains, and frames) are put to cognitive
work online in mental spaces in working memory.
This view enables us to take into account a variety of metaphor-related
mental activities that speakers engage in at this level. These include getting
primed to use particular metaphors by context, giving metaphorical expres-
sions specific socio-pragmatic functions, creating novel metaphors, using
metaphors deliberately, and several others.
Two quintessential online metaphorical phenomena were discussed in some
detail in the chapter: mixing metaphors and conceptual integration. I argued
that both assume and are heavily influenced by “higher” level image-schema,
domain, and frame metaphors. The analysis of mixing metaphors and concep-
tual integration indicated that, although much of the mental activity is
happening at the level of mental spaces in both cases, it would not be possible
to account for their use without taking into account the corresponding offline
conceptual structures. In the extended view of CMT, the offline and online
metaphorical structures are all needed for a (more or less) complete under-
standing of how conceptual metaphors work in natural discourse.
7 The Shape of the Extended View of CMT
In this next to last chapter, I attempt to identify the main ingredients of the
extended view of CMT, characterize the most important distinctions it
involves, outline the broad shape of the new view, and draw out its implica-
tions for a general understanding of metaphor. In the last chapter, I revisit the
five questions that inform the content and structure of the extended view, as
presented in this book.
7.1 Two Major Issues for CMT

To better understand and resolve some of the issues that beset CMT, let us
return to two of the examples (in the mountain stage and whisper) that have
been discussed in the book and approach them by asking two questions:
Why are these particular metaphors (unconsciously) chosen by the speaker in
the given context?
How can the speaker create (and the hearer comprehend) the particular
meaning expressed by the metaphors?
Why Are These Metaphors (Unconsciously) Chosen by the Speaker

in a Given Context?
We can begin with the metaphorical expression in the mountain stage in
relation to the “Armstrong doping scandal” case. Why was this particular
metaphor used by the crisis management expert? As was suggested in Chapter 4,
the specialist had extensive knowledge about the topic of the discourse, which
was Armstrong’s doping scandal. That knowledge included that Armstrong
participated in several Tour de France events and that this race has several
“mountain stages.” In other words, the topic of the discourse primed the speaker
to choose a metaphor to express a particular idea; namely, that, in order to come
completely clean, Armstrong had a long and difficult way to go.
As regards whisper, it was suggested in Chapter 5 that the metaphor
emerged from the physical(-perceptual) context in which it was produced.
Dr. Kumi Naidoo created the metaphor whisper against a background in which
150
7.1 Two Major Issues for CMT 151
there was a very loud rock concert and a comparatively quiet summit meeting.
The loudness and the relative quiet of the respective events were perceptual
features of the events. In this case, the original conceptualizer (Dr. Naidoo)
was prompted to use a perceptual property of the physical context from the
large body of experiential content that was available to him.
How Can the Speaker Create (and the Hearer Comprehend)

the Particular Meaning Expressed by the Metaphors?
Next, let us ask how the meanings of the metaphorical expressions were
created (and comprehended) by the speaker (and hearer). In the case of in
the mountain stage, the idea of a “long and difficult way to go” (i.e., in
Armstrong’s confessions) was expressed by the “mountain stage” metaphor.
Because of the wide conceptual gap between this meaning and the action is
motion metaphor, the following schematicity hierarchy was proposed in
Chapter 4 to bridge the gap:
Image schema level:
activity is motion
Domain level:
communication is self-propelled (forward) motion
Frame level:
confessions are races
Mental-space level:
confessions are bicycle races: the length and difficulty of
confessing wrongdoings is the length and difficulty of
being in the mountain stage of the tour de france
In other words, a schematicity hierarchy is established that serves as a concep-
tual pathway from the meaning expressed at the mental-space level all the way
to the image-schema level.
Let us now consider the case of whisper. Here as well, there seems to be a
great conceptual distance between the meaning “the lack of resolve and
effectiveness” (i.e., determination) and the meaning of whisper. In my book
Where metaphors come from (2015a: 58), I suggested we bridge the concep-
tual gap by a set of conceptual metonymies and a conceptual metaphor:
Metaphor:
intensity is strength of effect
Metonymies:
emotional responses for the emotions
emotion for determination to act
angry behavior for anger/argument
emotion for determination to act
Clearly, this is not a metaphorical schematicity hierarchy as discussed in this

book; we just have a single conceptual metaphor. What we have instead is a
pathway that is very different from the conceptual pathways I have proposed in
previous chapters. We can call it a non-systematic and spontaneous, or ad hoc,
conceptual pathway. If correct, this observation shows that not all metaphor-
ical conceptual pathways are constituted by schematicity hierarchies of the
kind we have seen so far in the book.
But now I propose to reconsider this solution and entertain the possibility
that there is indeed a systematic schematicity hierarchy connecting the
meaning “lack of resolve” and the image-schema metaphor intensity is
strength of effect. The new proposal is based on the idea that whisper
assumes a domain-level conceptual metaphor involving the generic-level
concepts of political action and linguistic action. Political actions
are typically talked about and conceptualized as linguistic actions, such as
“political speeches,” “negotiations between two countries,” “the opening talk
of the demonstration,” “the politicians continuing their conversation,” “the two
countries not talking again,” “the haggling over territories,” “a politician
bargaining for something,” and many others. Such examples are not simply
about being engaged in talk of some kind (i.e., linguistic action), they are about
doing, or conducting, politics itself (i.e., political action). If this is the case, we
are dealing with the domain-level conceptual metaphor: political action
is linguistic action, which would have as one of its mappings (the
relevant one): the intensity of political action is the strength
of effect of the linguistic action. A more specific frame-level
metaphor would be doing politics is speaking. Within this frame, one
mapping involves “determination, or resolve, to make changes.” This further
specifies the domain-level metaphor to: the intensity of political
determination to make changes is the degree of the volume
of speaking, or simply, political determination is loudness of
speaking. This last frame-level conceptual metaphor is further elaborated at
the mental space level, with the intensity set to “low,” both for the target and
the source. In other words, we can see these conceptual metaphors as consti-
tuting a schematicity hierarchy as follows:
Image schema level:
intensity is strength of effect
Domain level:
political action is linguistic action: the intensity of political
action is the strength of effect of the linguistic action
Frame level:
doing politics is speaking (to/with others): the intensity of
political determination to make changes is the degree of
the volume of speaking
7.1 Two Major Issues for CMT 153
Mental space level:

the low level of political determination to make changes is
the low level in the volume of speaking
The potential theoretical significance of how we decide on the analysis of
whisper is whether we have only one type of conceptual pathway or two for
conceptual metaphors. The second option in the analysis of whisper above
would support the existence of only one type of conceptual pathway, that of a
schematicity hierarchy.
However, if we accept the first option for the analysis of whisper above, it
might be the case that two types of conceptual pathways can be distinguished:
schematicity hierarchies (e.g., for in the mountain stage, option
two above)
spontaneous mixed pathways (e.g., for whisper, option one above)
The two conceptual pathways could be regarded as producing two types of
metaphors:
systematic metaphors (based on schematicity hierarchies)
non-systematic (isolated) metaphors (based on spontaneous mixed
pathways)
In other words, the suggestion is that examples of metaphors that lend them-
selves to be characterized by schematicity hierarchies form large systems of
conceptual metaphors and metaphorical expressions on multiple overarching
levels (mental space, frame, etc.), whereas examples of metaphors that do not
evoke schematicity hierarchies are isolated metaphorical linguistic expres-
sions. The latter possibility does not mean, however, that isolated metaphors,
such as whisper, are created (and comprehended) without an appropriate
conceptual pathway. They do require a conceptual pathway, but the pathway
will be of a spontaneous, or ad hoc, and, possibly, mixed type.
In addition to these two types of conceptual pathways, there is a third one, in
which two domains share a single image schema. This is not an image schema
metaphor with two distinct frames (as in the cases above), but simply a single
image schema that characterizes metaphorical analogies (or, more generally,
resemblance metaphors). The image schemas that are shared can be conven-
tional, entrenched ones (container, motion, link, etc.) or unconventional,
novel ones.
To take an example for the latter, consider a case of metaphorical conceptu-
alization I analyzed in previous work (Kövecses 2015a: 104–105). Frank Jump
is a New York City-based photographer who has AIDS and, as an artist, he
photographs old New York City mural advertisements. In a meeting with his
mentor, he talks with him about his art. This is how Jump reports on his
conversation:
In the beginning, I didn’t make the connection between the subject matter and my own
sero-positivity. I was asked to be part of the Day Without Art exhibition a few years ago
and didn’t think I was worthy – other artists’ work was much more HIV-specific. . . .
But my mentor said, “Don’t you see the connection? You’re documenting something
that was never intended to live this long. You never intended to live this long.” [italics
in the original]
The mentor calls Jump’s attention to the metaphorical relationship between his
illness and his art. We can state this in the following metaphorical analogy:
surviving aids despite predictions to the contrary is for the
old mural advertisements to survive their expected “life
span.” The two domains share an image schema; namely, that there is an
entity that exists longer than expected. The image schema is novel and
unconventional.
If this is on the right track, we can have three (not two) conceptual pathways
for the creation and comprehension of conceptual metaphors including meta-
phorical analogies:
schematicity hierarchy pathways
ad hoc pathways
shared image schema pathways
The third kind of pathway could serve as the basis for conceptual metaphors
that are usually regarded as metaphorical analogies. CMT research typically
foregrounds the types of metaphors that are based on schematicity hierarchies,
but there appears to be a large number of other cases that require ad hoc and
shared image schema pathways. In addition, in the section on the kinds of
metaphors in this chapter, I propose yet another conceptual pathway – that of
feature resemblance. Future research can quantify the precise proportion of the
different types.
7.2 Embodied Metaphor versus Discourse Metaphor

In a recent paper, Gibbs (2017c) distinguishes between “embodied” versus
“discourse metaphors.” Similarly, he distinguishes the theories used to account
for them: the embodied metaphor view used in cognitive linguistics and the
discourse views advocated mostly by more socially oriented metaphor
researchers. There are several theories of metaphor that attempt to account
for metaphor use in discourse, such as in the mountain stage and whisper.
Gibbs (2017c) discusses four of these in his paper: Zinken et al.’s (2008) view
of discourse metaphors, Quinn’s (1991) view that metaphors are simply
reflections of literal cultural models, Cameron’s (2008, 2010) discourse
dynamics approach to metaphor, and Steen’s (2008, 2011) theory of deliberate
metaphor. What these different views share is that they downplay the role or
7.2 Embodied Metaphor versus Discourse Metaphor 155
even reject the existence of conceptual metaphors of the kind mostly discussed
in the CMT literature (like life is a journey and the primary metaphors that
give them their embodied character) and, at the same time, they emphasize the
role of context in the creation and use of metaphors.
In his paper, Gibbs criticizes the four theories of discourse metaphor and
offers a synthesis of the discourse and embodied metaphor views in an
integrated model of “metaphor performance”: the dynamic systems view of
metaphor (see, e.g., Gibbs and Cameron 2007; Gibbs 2017c). I believe that the
dynamic systems view bears many similarities to the contextual and multilevel
view of metaphor that I am developing here. Both views share the goal of
characterizing metaphor performance (just like the discourse metaphor views),
but they do so by taking advantage of (rather than rejecting) embodied
conceptual metaphors – though to different degrees. In the remainder of the
section I briefly outline how my proposed account could achieve this. To
demonstrate, I use, in addition to the metaphors in the mountain stage and
whisper, the metaphor capsize, as it was used in a piece of natural discourse in
a 2007 newspaper article I analyzed in previous work (Kövecses 2015a).
Following hurricane Katrina that devastated New Orleans in 2005, a jour-
nalist did an interview with Fats Domino, a now elderly American rock
musician living near New Orleans. The journalist writes:
The 2005 hurricane capsized Domino’s life, though he’s loath to confess any incon-
venience or misery outside of missing his social circle . . . (USA TODAY, 2007,
September 21, Section 6B) (taken from Kövecses 2015a: 103).
The interview was largely about how Domino’s life was affected by the
hurricane. Its very negative effect is expressed by the unconventional linguistic
metaphor capsize, which is an instance of the applicable conceptual metaphor:
life is a (sea) journey. At the time of the interview, there were still many
noticeable signs of the devastation, as the newspaper article about domino
indicates. The journalist continues the quote above as follows: “. . . in the
Lower Ninth Ward, still destroyed and deserted despite some signs of
renewal.”
I suggest that this salient perceptual property of the physical situation plays a
major role (but likely not the only role, as suggested to me by Michele Feist in
personal communication, March 15, 2019) in the metaphorical conceptualization
of Domino’s life: the still visible sight of the devastation caused by the hurricane
(situational context), the memory of the overturned boats (conceptual-cognitive
context), the topic of the conversation with Domino (discourse context), and the
universal experience of falling down and not being able to function (bodily
context) prime the conceptualizer (the journalist) to use the verb capsize, which
assumes (and evokes) the primary metaphors of action is self-propelled
motion and persisting or functionality is remaining erect.
Simultaneously, since the topic of the interview is Domino’s life, the life and
travel domains are evoked, resulting in the conceptual metaphor life is
travel. An aspect of this metaphor is the correspondence between leading
one’s life and journeying, that is, the metaphor living a life is journey-
ing. (As noticed by Framenet, journeying is the activity aspect of travel,
which is a more inclusive category than journey.) And since the journalist’s
immediate idea at this point in the discourse concerns how Domino’s life has
suddenly changed for the worse, a very specific metaphor is evoked – that of
a sudden, unexpected turn of events for the worse in domi-
no’s life is the capsizing of domino’s boat in the course of his
sea journey, as expressed by the verb capsize.
In other words, primed by the context, a schematicity hierarchy is con-
structed online, which can be represented as follows:
Image schema:
action is self-propelled motion + persisting / functionality
is remaining erect
Domain:
life is travel
Frame:
living a life is journeying
Mental space:
a sudden, unexpected turn of events for the worse in
domino’s life is the capsizing of domino’s boat in the
course of his sea journey
The conceptual metaphor at the mental-space level is detailed, rich, and

specific. It is essentially an explication of the metaphorical meaning of capsize
in the discourse in which it occurs. This meaning is, using traditional semantic
terminology, both denotative and connotative. (I will explain the relevance of
this distinction below.)
In sum, the metaphorical conceptualization that is expressed in the phrase
“capsized Domino’s life” results from the priming effect of one or several
contextual factors that trigger the setting-up of a schematicity hierarchy online
in the situation of discourse. This way, the proposed extended CMT success-
fully combines the major features of the embodied and discourse metaphor
views: embodiment as based on experiential correlations and context, respect-
ively. But of course this is only applicable to those metaphors that are
embodied in the sense of standard CMT view (only they can have a schema-
ticity hierarchy associated with them). What about those metaphors that are
viewed by several researchers as “discourse” metaphors but that are based on
resemblance, rather than correlation? I will come back to this issue in another
section below.
7.3 Kinds of Metaphorical Meaning 157
7.3 Kinds of Metaphorical Meaning

Given the metaphorical expressions we have considered so far, it seems
appropriate to distinguish three types of metaphorical meaning: meaningful-
ness, decontextualized meaning, and contextualized meaning. The distinctions
are based on the schematicity hierarchy and my earlier distinction between the
subindividual, supraindividual, and individual levels of metaphor (see
Kövecses 2002/2010).
Meaningfulness is based on the image schema level, where image-schema
metaphors provide naturalness and bodily motivation for particular cases of
metaphorical conceptualization. This is achieved by means of highly sche-
matic metaphors – often, primary metaphors, although, as Grady (1997a)
notes, not all image-schema metaphors are primary metaphors (but many are,
e.g., states are containers, more is up, action is motion) and not
all primary metaphors are image-schema metaphor (e.g., purposes are
destinations). Several primary metaphors are at the level of domains or
frames (such as understanding is seeing, purposes are destin-
ations). Image schemas constitute the level of meaning that in previous
work I called the “subindividual level” (Kövecses 2002/2010). But the main
point here is that image-schema metaphors make particular metaphor hierarch-
ies embodied, and hence lend them a feel of naturalness. In other words, they
make them meaningful.
Decontextualized meaning is based on the domain and frame levels. Here,
meanings are based on more specific conceptual metaphors (than the image-
schematic ones) and the mappings that constitute them. The elements of two
domains or frames are in a conventional relationship. These are the mappings,
or correspondences, that define the conventionalized meanings of metaphorical
expressions (e.g., “heat ➔ intense anger” defines “intense anger” as the
meaning of burning with anger). Very often, such conventionalized meanings
are lexicalized for the source domain term (e.g., burn). Decontextualized
meaning is, therefore, located at the “supraindividual level” and it is shared,
to a large extent, across individuals in a speech community (Kövecses 2002/
2010). Decontextualized meaning corresponds to what I called denotational
meaning above.
Contextualized, or rather, contextual, meaning can be found on the “indi-
vidual level” (Kövecses 2002/2010). This is the level that corresponds to the
level of mental spaces in metaphorical schematicity hierarchies. At this level,
the decontextualized structures and the decontextualized meanings that rest on
them become fully individuated, specific, detailed, and rich in conceptual
content. This can happen in large part due to narrowing down full-blown
domain and frame-level structures to one or a few of their aspects and
expanding on these aspects in several ways. This is what we saw above in
the example of capsize. Capsizing is just a small part of the sea journey
frame. Its contextual meaning is “changing for the worse.” But it (capsizing)
has its own conceptual potential to expand by means of drawing out its
implications, such as one’s life being normal before it happens, the change
being sudden and unexpected, uncertainty about the future following the
change, the change for the worse evoking empathy and compassion, and
others. In a way, it can be suggested that a particular contextual meaning is
introduced in order to enable a variety of social, pragmatic, emotive, rhetorical,
etc., functions and effects. In traditional terminology, this is called connotative
meaning. Contextual meaning would, then, correspond to connotative mean-
ing, mentioned above.
It is this type of meaning construction that is the true “element” of concep-
tual integration; it is here that conceptual integration can break up conventional
conceptual structures, reassemble them in new ways, and can construct novel
meanings. And the same applies to deliberate metaphors that are based on
conceptual metaphors and come with a schematicity hierarchy. However, this
is not to claim, of course, that either blending or deliberate metaphors can only
rely on conceptual metaphors (those based on correlations in experience). But
in the case of conceptual metaphors based on correlations in experience, the
three kinds of meaning are jointly present in every use of a metaphorical
expression.
7.4 Kinds of Metaphors

Several researchers have pointed out that the kind of metaphors most com-
monly studied by CMT scholars (i.e., conceptual metaphors based on experi-
ential correlations) are not the only kind of metaphor (see, e.g., Evans 2013;
Grady 1999). For example, Grady suggested that there are essentially two
kinds of metaphor: those that are based on correlations in experience (concep-
tual metaphors “proper”) and those that are based on some kind of resemblance
(resemblance metaphors). Evans (2013) distinguishes conceptual metaphors
from what he calls “discourse metaphors.” The latter include image metaphors
(e.g., Andre Breton’s poetic metaphor “My wife whose . . . waist is an hour-
glass” in Lakoff and Turner 1989), metaphors with a predicate nominative
form (e.g., “My job is a jail”), and examples like a credit card tart. This
classification partially overlaps with Gibbs’ category of discourse metaphor;
namely, both include what Zinken (2007) calls discourse metaphor. Evans uses
the example credit card tart for Zinken’s narrower conception of discourse
metaphor. This last metaphor is a recent development in British English (Evans
2013). Tart here means a “person whose fidelity is unreliable in any sphere,”
including banking habits, which developed from an earlier sense of the word,
namely, “prostitute.” Common to discourse metaphors is that they are based on
7.4 Kinds of Metaphors 159
some kind of resemblance. Also, they characteristically start out as novel

metaphors, which either drop out of use or remain in the language in the form
of lexicalized units. If the latter, they often acquire more general meanings,
such as the cases of tart, in line with Bowdle and Genter’s Career of Metaphor
Hypothesis (Bowdle and Gentner 2005).
The groupings of metaphors by Grady and Evans basically boil down to a
two-way distinction: conceptual metaphors, which are based on correlations in
experience, and discourse metaphors, which are based on some kind of
resemblance. The question now is what the multilevel and contextual view
of metaphor can contribute to the classification of metaphors. In order to see
this, let us examine the case of the metaphor whisper again.
In essence, the extended view of CMT confirms the two-way classification if
we accept the second, new analysis of whisper above as valid (in which it is
based on a schematicity hierarchy). But if we accept the 2015 account (as
given above) as valid (in which the meaning of whisper is based on an ad hoc
conceptual pathway), then we have a somewhat different classification on our
hands. Let us take a look at the latter view.
We could say that, unlike discourse metaphors, whisper is not based on
resemblance. It is difficult to see a shared feature or structure between whisper
and “lack of resolve, or determination.” But, we could continue to argue that,
unlike (instantiations of ) conceptual metaphors proper, whisper is not based
on a correlation in experience either. Although we have the intensity is
strength of effect metaphor in the conceptual pathway of whisper, there
do not appear to exist any corresponding and matching domain- and frame-
level conceptual metaphors that would accommodate whisper in its metaphor-
ical sense as one of their linguistic examples. If this observation is valid, there
is no schematicity hierarchy that conceptualizers could (unconsciously) build
either in order to create or make sense of the word in the discourse in which
it occurred.
But, then, if there is no schematicity hierarchy and there is no resemblance
on which the meaning could be based, how can people produce and compre-
hend it as naturally as they do? My response was that they can build a
spontaneous mixed conceptual pathway to figure out what the word means
in context. In the case of whisper, the pathway includes a generic, image-
schema level metaphor and a number of metonymies that lead to the intended
contextual meaning. If correct, the implication of this analysis would be that
there is a third kind of metaphor, in addition to correlation- versus
resemblance-based ones: a kind that is created and comprehended by means
of an ad hoc conceptual pathway, such as we have seen for whisper.
On this analysis of whisper, a spontaneous (mixed) conceptual pathway is
created that results in a contextual meaning. We could characterize the general
case (for metaphors like whisper) as follows:
mental-space level metaphor

"
AD HOC CONCEPTUAL PATHWAY (including metonymies)
Image schema metaphor
"
CORRELATION IN EXPERIENCE
In other words, we have an isolated correlation metaphor (intensity is

strength of effect) without corresponding and matching domain- and
frame-level metaphors. This indicates that the mere existence of a correlation
does not guarantee the emergence of a schematicity hierarchy.
But if the new analysis for whisper (see above) is accepted as valid, we
would then have the following general picture for conceptual metaphors (of the
correlation kind). In this case, an entire schematicity hierarchy is utilized:
"
frame-level metaphor
domain-level metaphor
image-schema-level metaphor
"
CORRELATION IN EXPERIENCE
Here, there are multiple mappings at the schema, domain, and frame levels,
such that ELEMENT A in a source frame, domain, and schema maps onto
ELEMENT A in the target frame, domain, and schema, ELEMENT B in the
source frame, domain, and schema onto ELEMENT B in the target frame,
domain, and schema, and so on for ELEMENTS C, D, E, etc. on all three
levels. The fourth level, that of mental space, is dedicated to a single mapping,
where the metaphorical contextual meaning is conveyed.
And in the case of one class of resemblance metaphors, there is a single
mapping from the source entity to the target entity, which exploits a single
feature:
"
ENTITY 1: FEATURE A – ENTITY 2: FEATURE A
"
FEATURE RESEMBLANCE
An example for this case could be: “John is a pig,” where an assumed feature
of pigs (being dirty, messy) is attributed to John (by way of a metonymic
process: the whole entity for a feature that characterizes
that entity). This is essentially a situation that operates with an (assumed)
analogy (resemblance) between features of two different entities. (Moreover,
the assumed analogy between the two entities is based on the prior personifi-
cation of pigs, that is, on the metaphor pigs (animals) are people.)
“Feature resemblance” can be seen as a further type of conceptual pathway
besides schematicity hierarchy pathways, ad hoc pathways, and shared image
schema pathways. This type of pathway produces non-systematic (isolated)
linguistic metaphors, such as pig in the example above.
But analogies are usually seen as multiple mappings between two frames or
domains. Consider the expression gut the building, as used in the following
example:
Kirk E. Kreuzwieser, vice president of Strollo Architects, says much of the interior of
the 96-year-old Wells Building needs to be gutted. He and Gregg Strollo, company
president, believe the $4 million-plus rehabilitation project could be completed in less
than nine months after financial issues are resolved. (www.vindy.com/photos/2013/jun/
17/68965/)
This is a domain- or frame-level metaphorical analogy that is constituted by
multiple mappings (Figure 7.1). The mappings are:
the person who disembowels a dead animal ➔ the person who removes the
interior of a building
the animal body ➔ the building
the bowels ➔ the interior
the action of disemboweling ➔ the action of removing the interior
The mappings are based on a structural analogy between two frames or
domains, which means that the propositionally-explicated mappings between
the two frames or domains (here, animal and building) share an image-
schematic structure, an image schema, something like this:
Source Target
ANIMAL BUILDING
Figure 7.1 Shared image schema
Thus, structural analogy (or resemblance) means that the source and the
target share some image schema (here, internal container (or object) moved out
of external container). Of the four mappings above, the last one appears at the
mental-space level. This kind of metaphor could be represented as follows:

"
frame-level metaphor
or
domain-level metaphor
"
ELEMENT A – ELEMENT A
ELEMENT B – ELEMENT B
ELEMENT C – ELEMENT C
etc.
"
STRUCTURAL RESEMBLANCE (shared image schema)
In this case, unlike correlation metaphors, there is a shared image schema (not
an image-schema metaphor) and there is only a single set of mappings between
the elements – either on the frame or domain level (instead of the three sets of
mappings at the image schema, domain, and frame levels). At the mental-space
level, the mapping will be limited to one, which conveys the intended context-
ual meaning. This is the general representation for cases like gut the building
and also for cases like the Frank Jump example discussed above in the chapter.
If we employ a broader definition of conceptual metaphors as any case of
multiple mappings from a source to a target (and not just those that are
correlation-based), many metaphors that are resemblance-based (those that
map entire structures, and not just a single feature) would also qualify as
conceptual metaphors (such as the metaphor above or, say, the an atom is
the solar system metaphor). These are usually called metaphorical analo-
gies. But, when they talk about conceptual metaphors, cognitive linguists
typically mean correlation-based metaphors, and, as we saw throughout, these
come with schematicity hierarchies with four levels (as discussed above).
A Note on “Cultural Metaphor”

The distinction between correlation and resemblance metaphors has an import-
ant consequence; namely, that the distinction can help us avoid debates
concerning the degree of the universality versus culture-specificity of concep-
tual metaphors. It has been noticed by several scholars even within the CMT
framework (e.g., Gibbs 1999a,b; Kövecses 2005) that, especially in the first
two decades of CMT research, much more emphasis was placed on the
universal aspects of conceptual metaphors than on their culture-specific
aspects. Most recently, Musolff (2016) also observes that in CMT
“. . . primary,” i.e., experience-based and neuro-physiologically manifested
metaphor-constitutive “image-schemas” are seen as “embodied conceptual
universals” (Gallese and Lakoff 2005; Lakoff 1993, 2008, Lakoff and Johnson
1980, 1999). Musolff opposes this strand of research to work on metaphor
from a cultural linguistic perspective (see, Sharifian 2011, 2015, 2017a,b),

where metaphors arise from cultural conceptualizations. To demonstrate,
Musolff cites his own research on whether or not people in different lan-
guages/cultures conceive of the notion of “body politic” (the idea that
members of a nation conceptualize their nation as the human body) in the
same way, that is, in a universal manner. His results indicate that it is not
the case; people from different nationalities view the role of the human body in
the conceptualization of “nation” in systematically different ways. He found
that the participants of the study interpret the nation-as-body metaphor in
several distinct ways (i.e., nonuniversally), including the most basic distinction
between the nation as an anatomy-/function-based entity (mostly for Western-
ers) and geography-based entity (mostly for Chinese participants). Musolff
suggests that these (and the other) interpretations can be attributed to different
culture-specific traditions (like those found in the West and China). In other
words, he shows that there is no automatic, universal response in the interpret-
ation of a conceptual metaphor that is expected to trigger automatic and
universal responses.
What does this issue have to do with the correlation versus resemblance
metaphor distinction? I believe that when leading representatives of CMT
(such as Lakoff and Johnson) talk about conceptual metaphors, they primarily
mean correlation-based metaphors – conceptual metaphors that come with a
schematicity hierarchy, as described in this book. Such conceptual metaphors
(because of their image-schematic basis in correlations in experience) tend to
be universal. By contrast, I do not think that the nation-as-body conceptual
metaphor is a correlation metaphor; it is a resemblance-based one. The human
body is not correlated in any way with the concept of nation, but it can bear
some generic similarities to it. Respondents may compare the nation to the
hierarchical functional organization of the human body (e.g., when the head is
seen as dominating the arms, legs) or to the arrangement of the body organs (in
which, e.g., the mouth and nose correspond to particular geographic locations
of the nation). In other words, keeping the correlation metaphors distinct from
the resemblance ones, we do not have to claim that both of these metaphor
types produce universality (only the correlation ones do) or that the correlation
metaphors cannot produce nonuniversality (though it is mainly the resem-
blance metaphors that do).
I say “mainly” because of course to some extent the schematicity hierarchies
can also produce nonuniversality, as I have shown in several places in this
book. The degree of nonuniversality increases as we move down the hierarchy
(from image schema metaphors to mental space metaphors) and it decreases as
we move up the hierarchy of levels. To put it differently, the degree of
universality is highest at the level of image schemas and lowest at the level
of mental spaces. The reason for this is that mental space level metaphorical
conceptualization is deeply embedded in context and is exposed to contextual

influence. This is the true level of “cultural conceptualization.” Such context-
ual influence equally affects correlation metaphors (schematicity hierarchies)
and resemblance metaphors (with shared image schemas).
7.5 The Shape of the Contextual and Multilevel View of CMT

The vision of CMT I am describing here builds on a number of distinctions
concerning meaning, conceptual structure, memory, ontological level, and
context. In this section, I would like to briefly review these, and then suggest
a general outline of how they are related to one another.
I have already treated the notion of meaning above in the chapter. All I need
to say at this point is that my proposal was that in metaphorical meaning
making three different types of meaning can be distinguished: meaningfulness,
decontextualized meaning, and contextual meaning. They all participate in the
use of conceptual metaphors proper.
As regards the conceptual structures (see Chapter 4 and this chapter)
that make up conceptual metaphors (image schema, domain, frame, mental
space), they are all constitutive parts: a conceptual metaphor consists of
image-schema, domain, frame, and mental space metaphors. They form an
interlocking hierarchy that is based on decreasing (or increasing) degrees of
schematicity – with the endpoints of image schema and mental space meta-
phors. Image-schema level metaphors (including primary metaphors) define
the generic-level meaningfulness of a conceptual metaphor. Domain- and
frame-level metaphors (with their mappings) are responsible for the decontext-
ualized meanings that are associated with them and are based on them. Mental
space metaphors allow for and capture the contextual meanings in a piece of
ongoing discourse.
The issue of memory is relevant and important because the different meta-
phors are located in different types of memory: long-term and working
memory. Long-term memory stores image-schema-, domain-, and frame-level
metaphors. Working memory is the site where mental-space-level metaphors
occur. Moreover, image-schema metaphors are stored in long-term memory in
the form of analogous structures, as opposed to the other three metaphors
(domain, frame, mental space) that have a propositional format. Finally, it is
only working memory that can give rise to conscious and deliberate metaphors
(á la Steen).
In previous work (Kövecses 2002/2010 and Chapter 4), I distinguished
various ontological levels where conceptual metaphors can exist: the subindi-
vidual, supraindividual, and individual levels. The subindividual level is the
ontological level where meaningfulness resides in long-term memory in
the form of image schema metaphors. At the supraindividual level, we find
7.5 The Shape of the Contextual and Multilevel View of CMT 165
Table 7.1 Summary of key ideas of the extended CMT view
Kind of Conceptual Ontological

Kind of Meaning Structure Kind of Memory Level
Meaningfulness Image-schema Long-term memory Subindividual

metaphor (Analog structure) level
Decontextualized Domain metaphor Long-term memory Supraindividual
meaning Frame metaphor (Propositional structure) level
Contextual meaning Mental-space Working memory Individual level
metaphor (Propositional structure)
Table 7.2 Four context types and their contextual factors
Situational Conceptual-Cognitive
Context Discourse Context Bodily Context Context
Physical Surrounding discourse Correlations in Metaphorical

environment (co-text) experience conceptual system
Cultural situation Previous discourse Bodily conditions Ideology
Social situation Knowledge about speaker, Body specificities Concerns and interests
topic, hearer
Dominant forms of History
discourse
domain- and frame-level metaphors in long-term memory with their decon-

textualized meanings. At the individual level, mental-space metaphors are used
to generate contextual meanings in working memory.
We can summarize these key aspects of the extended view as discussed so
far, together with their interrelations, in the table above (Table 7.1)
In addition to the above cognitive characteristics of conceptual metaphors,
there is another major aspect of the view I am proposing; it is the notion of
context. The extended CMT I envision does not only have a rich cognitive part,
as specified above, but it also has a rich contextual component (see Kövecses
2015a and Chapter 5). Admittedly somewhat radically, I distinguish four types
of context: the situational, discourse, bodily, and conceptual-cognitive context.
Each one of these comes with a number of empirically established contextual
factors. (For a detailed description, see Kövecses 2015a.) Table 7.2 shows the
major contextual factors I have found in real discourse that constitute the four
context types.
The main suggestion concerning these context types and contextual factors
was that any one of them can prompt a speaker to use a particular linguistic
metaphor that matches a contextual factor; that is, the idea was that they all
have a potential priming effect. A main issue I discussed in this book was how
the cognitive aspects and the contextual aspects of metaphors can be put
together in a coherent framework. This issue requires the formulation of a
hypothesis to explain how the speaker comes up with a metaphor to begin with
and how the hearer can understand this metaphor. At the beginning of the
present chapter, I looked at some examples of how these metaphor processes
can be accomplished within my extended CMT framework. No matter how
oversimplified and naïve it may be, I now attempt to turn the previous
observations about the cognitive and contextual aspects of conceptual meta-
phors into a process view, that is, into a dynamic model of mental processes in
the course of the construction of metaphorical meaning.
7.6 A Very Tentative and Informal Process Model

Imagine a discourse situation in which there is a speaker and a hearer. Since
they are engaged in discourse, there is a discourse situation, with a speaker, a
hearer, and a topic, in which they may share a considerable body of infor-
mation about all of these elements. (DISCOURSE CONTEXT) The speaker
and hearer have a great deal of physical, social, and cultural information in and
about the situation of discourse, much of which they share. (SITUATIONAL
CONTEXT) They have bodies, and they assume that their bodies function in
the same way, bringing some of their bodily experiences into correlation with
other concepts. But they also have knowledge about their own and the other’s
bodily conditions and bodily specificities. (BODILY CONTEXT) And finally,
to a large extent, they share a metaphorical system, they may subscribe to some
ideology in the culture, they may have certain interests and concerns in
life, and they may share some group or personal history. (CONCEPTUAL-
COGNITIVE CONTEXT)
At a point in discourse, the speaker’s communicative intention appears to be
best expressible by metaphor. The speaker (unconsciously) forms a contextual
meaning that is supported by an appropriate image-schematic metaphor. The
contextual factors mentioned above are all active in the speaker’s working
memory. One of the factors primes the speaker to build a conceptual pathway
from the contextual meaning as expressed by a conceptual metaphor at the
mental-space level that is supported by the image-schema metaphor. The
pathway is completed through the specification of the image-schema metaphor
at the frame- and domain-levels.
Up to this point, only the content part of the contextual meaning is created.
However, in many cases this content part is created in order to achieve some
simultaneous pragmatic effect. The pragmatic effect and the conceptual con-
tent emerge jointly – one in the service of the other. Humor, emotion, narrative
function, rhetorical effect, etc., can be just as prominent parts of metaphorical
7.7 Conclusions 167
THE EMERGING CONTENT PART OF THE METAPHOR IS SUPPLEMENTED BY A

PRAGMATIC FUNCTION OR EFFECT
↑
THE APPROPRIATE METAPHORICAL VEHICLE FOR THE CONTEXTUAL
MEANING EMERGES ON THE BASIS OF THE SCHEMATICITY HIERARCHY
↑
A SCHEMATICITY HIERARCHY IS ESTABLISHED AS A RESULT OF THE
INFLUENCE OF CONTEXTUAL FACTORS
↑
AN IMAGE SCHEMATIC METAPHOR IS IDENTIFED THAT SUPPORTS THE
CONTEXTUAL MEANING
↑
THE NEED FOR THE EXPRESSION OF A PARTICULAR METAPHORICAL
CONTEXTUAL MEANING EMERGES
↑
A LARGE BODY OF CONTEXTUAL FACTORS IS ACTIVE IN WORKING MEMORY
↑
PRESENCE OF DISCOURSE CONTEXT - SITUATIONAL CONTEXT - BODILY
CONTEXT - CONCEPTUAL-COGNITIVE CONTEXT
Figure 7.2 A process model for conceptual metaphors
contextual meaning as the conceptual content that “carries” this function or

effect. In other words, our informal account of the metaphor process would
look something like Figure 7.2 in a sketchy format.
I admit that this “model” for the production of metaphors is my “personal”
folk theory (if there is such a thing). All I wanted to point out with it is what,
I think, the main ingredients of metaphorical meaning construction are in the
case of conceptual metaphors and how these ingredients can be put together in
some, hopefully viable, logical arrangement. The upward arrows do not really
display temporal sequence; they indicate some logical order that can be
assumed in the process. The model is not intended to be a “replica” of the
online process that is really the case. I do not want to (and cannot) take on the
job of psycholinguists and experimental cognitive psychologists (not to men-
tion neuroscientists). However, I hope that my naïve theory can provide these
researchers with some useful ideas in their own work on metaphor production
and comprehension.
7.7 Conclusions
In the chapter, I have outlined the main components of, and their interrelations
in, an extended CMT. The new view builds especially on the multilevel nature
and the contextual embeddedness of conceptual metaphors.
As regards their multilevel nature, conceptual metaphors involve a variety of
components. These were identified as certain kinds of metaphorical meaning
(meaningfulness, decontextualized meaning, and contextual meaning), certain

kinds of conceptual structures (image schema, domain, frame, and mental
space), certain kinds of memory (long-term and working memory), and certain
kinds of ontological status (supraindividual, individual, and subindividual).
As regards contextualization, the extended view involves four types of
context and the contextual factors they subsume: Situational context: physical,
social, and cultural; Discourse context: surrounding discourse, previous dis-
course, knowledge about main elements of discourse, and dominant forms of
discourse; Bodily context: correlations in experience, bodily conditions, and
body specificities; Conceptual-cognitive context: metaphorical conceptual
system, ideology, concerns and interests, and history.
In the new view, we can distinguish three types of conceptual pathways:
schematicity hierarchy pathways, ad hoc pathways, and shared image schema
pathways. The application of these different pathways results in different kinds
of conceptual metaphors: systematic metaphors (based on schematicity
hierarchy pathways), non-systematic (isolated) metaphors (based on ad hoc
pathways), and systematic resemblance (analogy) metaphors (based on shared
image schema pathways).
Finally, a tentative and informal metaphor processing model was proposed –
a model that is intended to map what is going on in the head of the producer of
a metaphorical expression in a piece of discourse.
8 By Way of Conclusion: Responses
to the Five Questions
I structured the content of this book according to five major issues that were
presented as five questions in the titles of five chapters. These were:
(1) The abstract understood figuratively, the concrete understood literally,
but the concrete understood figuratively?
(2) Direct or indirect emergence?
(3) Domains, schemas, frames, or spaces?
(4) Conceptual or contextual?
(5) Offline or online?
These are deliberately provocative titles that express extreme and simplifying
alternative views as regards CMT. In this final chapter, I would like to briefly
discuss them one by one. It will become clear, and I hope the relevant chapters
make this evident, that I do not subscribe to any such alternative views by
themselves. Reality is much more complex than what any of the alternatives by
themselves would represent.
8.1 The Abstract Understood Figuratively, the Concrete Understood

Literally, but the Concrete Understood Figuratively?
Let us begin with the most general issue – that of literality versus figurativity.
CMT builds on the idea that abstract concepts are conceptualized as concrete,
physical ones (but not the other way around). This is because the latter
(concrete concepts) can be understood in a literal way: through direct, physical
experience, while the abstract ones cannot. But we saw in Chapter 2 that the
extent of literally understood concrete concepts must be very small; many of
them are also understood metaphorically and metonymically for various
reasons. However, this does not prevent us from using these figuratively
understood concrete concepts as source domains of conceptual metaphors,
that is, to conceptualize abstract concepts. Many abstract target domains are
conceptualized as figuratively understood concrete source domains – either
etymologically in diachrony or in synchrony. This leads to a dilemma within
169
170 By Way of Conclusion: Responses to the Five Questions
the theory. If many source domains are in part figurative, then it is not the case
that abstractions are only conceptualized by means of literally understood
source concepts.
My suggestion to handle this apparent dilemma for the theory was that many
source domains have ontological content that is based on tangible experience
but they are also partially figuratively understood. Many concepts can be said
to have both ontological content and figurative construal as aspects of their
meaning. In the case of many concrete concepts the ontological content
dominates over the figurative construal aspect. In the case of abstract concepts
the reverse is the case: figurative construal dominates over ontological content
(Figure 8.1). This can be diagrammed as follows:
Concrete concepts Abstract concepts
figurative construal
FIGURATIVE
CONSTRUAL
ONTOLOGICAL
CONTENT
ontological content
Figure 8.1 Figurative construal and ontological content in concrete and

abstract concepts
Abstract concepts have very little ontological content, but a great deal
of figurative construal (i.e., they are for the most part figuratively under-
stood). But they do have some ontological content. This is what we call
the bodily basis, the embodiment of abstract concepts (such as the body
heat of anger).
It is interesting to note in this regard that several researchers consider
emotion concepts as somewhat different from other abstract concepts (see,
e.g., Altarriba et al. 2004; Borghi et al. 2017). They suggest that emotion
concepts are more embodied than other abstract concepts. I believe, however,
that this claim is too strong and needs to be qualified. It seems to me that it is
only prototypical emotion concepts (such as anger, fear, sadness, joy) that are
more embodied (that is, have a great deal of bodily basis). Nonprototypical
ones often lack embodiment completely – at least if we limit embodiment to
external signals of emotion states. (But as Anna Borghi remarked in personal
communication, e-mail message, February 26, 2019, we should consider not
only exteroceptive signals of embodiment but also interoceptive ones, which
would further refine the picture. This is a complex issue that I cannot go into
8.2 Direct or Indirect Emergence? 171
here.) On the “cruder” view of emotion embodiment, hope, friendship, respect,

contempt, and several others do not appear to be characterized by embodiment.
Instead, they are entirely figuratively construed. The conceptual metaphors
used for this purpose come from outside the typical metaphors that characterize
prototypical emotion concepts (see Kövecses 2000b).
Thus, in general, it appears that abstract concepts are somewhat like
concrete ones, in that they are embodied. But are concrete concepts some-
what like abstract ones? Given that concrete concepts can be figuratively
understood, many of them are like abstract ones, in that they are in part
figuratively understood. In conceptual metaphors we can rely on such
(partially) figuratively understood concrete concepts as source domains
because it is their physical ontological content that is dominant and fore-
grounded. This way, they can serve a cognitively useful function in the
metaphorical process.
8.2 Direct or Indirect Emergence?

As we saw above in the previous chapter, metaphor is a diverse category – with
several different kinds. Ever since the inception of CMT, the kind of metaphor
that enjoys the highest prestige and is most actively studied within the field is
what we call today correlation metaphors. (They were not called that in 1980.)
The reason for this is, I think, that correlation metaphors help us explain how
many abstract concepts emerge in an entirely natural and automatic way.
A key role in the explanation is played by Grady’s notion of primary
metaphor. Primary metaphors are the cornerstone of the standard view of
metaphor within a CMT framework. With its help it has become possible to
account for how large organizing and potentially universal metaphors like
life is a journey emerge and are routinely and automatically learned. We
absolutely need no instruction of any sort to learn and use them. No instruction
is necessary because the primary metaphors that constitute this complex
metaphor, such as purposes are destinations, difficulties are obs-
tacles, etc., are acquired as a matter of course, as we live our everyday
realities. For example, in many of our life experiences purposes are closely
associated with destinations and this correlation gives rise to the respective
primary metaphor.
But the question I raised in Chapter 3 was whether in the case of primary
metaphors these metaphors emerge directly from experience or not. Or to put
the same question differently, in the case of correlation metaphors is the
process of metaphorization itself a primary (and thus basic, foundational)
process or not? I argued it is not. I suggested that it must be preceded by
a metonymic stage that prepares the ground for the metaphorization
process. I showed that this can happen in essentially two ways. (1) In an initial
“frame-like” structure with element 1 (E1) and element 2 (E2), E1 and E2 are
correlated in experience, and thus E2 can stand metonymically for E1. For
instance, body heat stands metonymically for anger. Then, element 2 is gener-
alized (or schematized) and becomes the source concept for the initial concept,
which becomes the target concept. That is to say, body heat is generalized to
heat outside the emotion domain, thus allowing the primary metaphor anger
is heat to emerge. (2) In an initial “frame-like” structure with element 1 (E1)
and element 2 (E2), E1 and E2 are correlated in experience, and thus E2 can
stand metonymically for E1. Then, E2 is generalized and becomes the target
concept for the initial concept, which becomes the source. For instance, in the
initial “going to a destination” frame going to a destination involves the
purpose of arriving at that destination (thus, the two are correlated), which
gives us the metonymy arriving at a destination for achieving a
goal. But not all goals are related to arriving at destinations. Destination-
related goals are generalized to purposes in general. This yields the purposes
are destinations primary metaphor.
It is not my intention with this analysis to suggest that primary metaphors
are in any way unimportant. On the contrary, they do play the extremely
important role in human cognition in general and metaphorical conceptual-
ization in particular that was given to them in the theory of embodied cogni-
tion. My suggestion simply is that the metaphorization process that leads to
primary metaphors is preceded by a metonymization process, a metonymic
stage. In this sense, primary metaphors do not emerge directly. The metaphor-
ization process that leads to primary metaphors is not primary; it builds on the
prior process of metonymization.
In addition to its role in primary metaphors, metonymy also plays a part in
the formation of abstract concepts in general. But while in primary metaphors
metonymy assumes two closely correlated concepts, in the formation of
abstract concepts it can perform its role within a single concept, where the
original concept involves its own schematized copy. For example, the primary
meaning of the verb gut (“to take the interior organs of a dead animal out of
it”), analyzed in a previous section of this chapter, involves as its schematized
conceptual copy “to remove something out of something.” This schematization
works by metonymy, which, in this case, would be specific meaning for
its corresponding generic meaning or, more generally, specific for
generic. Thus, in the primary physical sense we have a human agent taking
out the internal organs of a dead animal body and in the abstract sense we
have an agent removing something out of something. The generic meaning
does not specify what is taken out, how it is taken out, and what it is taken out
of. This bleached version of the primary sense of gut is a schematized copy of
8.3 Domains, Schemas, Frames, or Spaces? 173
that sense, such that the specific meaning stands metonymically for the
generic one. As a result, an abstract sense of gut is created that is applicable
to several different domains such as economy or law (as indicated, e.g., by
Merriam-Webster).
It seems then that the metonymization process is crucial not only in creating
primary metaphors but also in the emergence of abstract concepts in general. In
previous work (Kövecses 2015a: 22–24), I noted “metaphorical schematization”
works by conceptual metonymies, and the process gives us a number of abstract
concepts based on metonymy, including such concepts as heaven, provi-
dence, and causation / creation.
8.3 Domains, Schemas, Frames, or Spaces?

Conceptual metaphors and the conceptual structures that they involve are
discussed under many different names. Lakoff and Johnson preferred to use
domain (as in one domain understood in terms of another domain), but they
also talked about experiential gestalts in their early work. Lakoff (1996)
mostly refers to conceptual metaphors as a relationship between two frames.
Lakoff and Kövecses (1987) use the term cognitive model, and many cultural
anthropologists called the same thing cultural model (Holland and Quinn
1987). The term schema also became popular in the wake of Lakoff and
Turner’s (1989) work, together of course with image schema that originated
in Johnson (1987) book. More recent discourse-oriented work (e.g., Musolff
2006) found the term scenario, or scene, as more appropriate in describing
metaphor-related phenomena in discourse. And researchers in conceptual
integrations theory adopted such designations to the conceptual structures
involved in metaphor as space, mental space or, even simply, input (space).
And the list could no doubt be continued.
Did all of these authors mistakenly use the designations they adopted? Not
at all, they all used them correctly. What unifies all the terms that became
available was a simple but powerful idea: we experience aspects of the world
in terms of coherent mental organizations of our experience. Domain, frame,
schema, model, scenario, space, and so on, all capture the gestalt character of
how we experience the world. Conceptual metaphor as a way of experiencing
the world can involve any one of these constructs.
It was recognized early on that conceptual metaphors can be stated at
various levels of generality (on this, see, e.g., Clausner and Croft 1997). For
example, Lakoff and Johnson (1980) observed that the time is money
metaphor can also be given as time is a resource, and this way more
general metaphorical expressions, such as use time or give time, can be
accommodated within the same metaphor. The resource metaphor is more
general or schematic than the MONEY metaphor. The question is whether

concepts like resource and money constitute domains or frames, or one is a
domain and the other is a frame.
What was initially less clear in connection with such gestalt-like concep-
tual structures used in metaphor was that the various conceptual structures
(designated by the mnemonics a is b, such as love is a journey) tend to
occupy different levels on a scale of specificity-schematicity. In his 1993 paper
Lakoff distinguished a number of conceptual metaphors in a hierarchy of
three tiers:
The Event Structure metaphor
life is a journey
love is a journey
He states that they are different in their degree of schematicity: that life events
are special cases of events in general, and the events in a love relationship are
subcases of life events. Lower level metaphors inherit higher level metaphor-
ical structures.
Dancygier and Sweetser (2014: 43–49) adds to this that at the level of the
Event Structure metaphor we are dealing with image schema frames (as they
refer to more complex image schemas) that can be elaborated by frames at
lower levels. For instance, the journey frame is a rich subframe of the
image-schematic frame directed motion on a path (composed of the
basic image schemas of motion, path, location). The frame of journey
can then be used to structure various domains, including life, relationships,
careers, projects, etc. – almost any long-term activity. What this means is
not only that the particular conceptual metaphors themselves can form
schematicity hierarchies, but also that the conceptual structures that constitute
these conceptual metaphors form such hierarchies, as I have demonstrated
throughout:
Image schema structures (most schematic)
Domain structures (less schematic)
Frame structures (less schematic)
We can call all of them frames or domains (in that they all represent coherent
organizations of experience), but when we see these conceptual structures (and
the corresponding metaphors) as being relatively more or less schematic, the
structures on top will always be image schemas, with domains below them,
and with frames below domains.
I noted in Chapter 4 that my view of schematicity hierarchies differs from
the standard view in two ways. The first is that in my conception the least
schematic conceptual structure in a hierarchy is not a frame but a mental space.
8.4 Conceptual or Contextual? 175
Mental spaces elaborate on frames, making them as specific as required by the

given discourse situation. In other words, I suggest a four-level hierarchy
(in addition to the level of linguistic expression):
Image schema structures (most schematic)
Domain structures (less schematic)
Frame structures (less schematic)
Mental space structures (least schematic)
(Linguistic expression)
Mental spaces usually foreground a single metaphorical mapping between two
frames, but this mapping, which expresses the contextual meaning at a given
point in discourse, evokes the entire hierarchy “above” it. This way, it can be
shown that the functioning of schematicity hierarchies is not limited to con-
ceptual structures in long-term memory, but is closely linked to how these
structures are mobilized in working memory.
The second way my treatment of schematicity hierarchies is different from
the standard view is that I see each and every hierarchy as functioning within a
rich context. Various aspects of the contexts (i.e., the contextual factors) can
have an influence on which conceptual metaphor (and the hierarchy that goes
with it) is utilized to conceptualize the situation. The same factors can also be
crucial in the (unconscious) choice of the linguistic metaphors that can best
express the conceptual metaphor in the situation of discourse.
8.4 Conceptual or Contextual?

A major discovery of CMT was that metaphors are not simply individual
linguistic expressions but systems of concepts in the form of mappings
between conceptual gestalts, such as domains, frames, schemas, and so on.
CMT thus made a natural connection between metaphor in language and
metaphor in thought (the metaphorical conceptual system). It maintains that
many metaphors in language only exist because we have conceptual
metaphors.
Actually, the notion of schematicity hierarchy further extends the view of
metaphor as a device in the way we think, and not as a rhetorical means in
language. The many conceptual metaphors we have are not independent from
each other, but form larger hierarchical systems that make metaphorical
thought extremely easy and efficient without conscious awareness. The sche-
matity hierarchies that we have seen in the book and the ones described in
work by Lakoff, Sweetser, Dancygier, and others show the robustness of the
phenomenon and the power it has on and in how we conceptualize the world
around us.
This large conceptual scaffolding of conceptual metaphors is formed,

informed, enriched, and shaped by our experiences. It is the totality of
our experiences that, to my mind, constitutes context (see Kövecses 2015a
and Chapter 5). In an effort to capture the idea that the totality of our
experiences can play a role in shaping the metaphorical system, I think of
context as composed of four types: situational, discourse, conceptual-
cognitive, and bodily. The four types each include a variety of more
specific contextual factors that can affect the formation and functioning of
the large metaphorical system with its many schematicity hierarchies. As
cognitive linguists and scientists have shown (e.g., Lakoff and Johnson
1999; Gibbs 2006), universal aspects of the human body play an especially
significant role in how this system is formed and how it functions. In
addition to the part played by primary metaphors, other image schema
metaphors and the metonymies that enable the formation of primary meta-
phors are crucial in this regard.
But the body is not only universal – it is also, often, nonuniversal; it can
have its own specificities and, sometimes, very unique conditions. As a result,
the body can also produce metaphors that are specific to groups of people or
even individuals. For this reason, I find it legitimate to regard the body as a
kind of context that can shape the metaphor system both in universal and
nonuniversal ways.
What I call the conceptual-cognitive context includes the cultural factor of
“interests and concerns.” These apply to both groups and individuals. For
example, if in a society people are very interested in sports, the domain of sport
will be commonly used in schematicity hierarchies where competition and
control are an issue.
In an even more obvious way than the previous two, the situational and
discourse contexts exert an influence on how we conceptualize the world
metaphorically. We have seen a number of examples for these in this chapter
and in Chapter 5.
In general, we can suggest that contextual influence is strongest and most
immediate at the mental spaces level in particular local discourse situations. In
the case of frame and domain level metaphors, contextual influence is not
direct and immediate. These metaphors appear to be stable in given speech
communities and change over longer periods of time as a result of changes in
context. (On this, see, e.g., Trim’s work on the evolution of several conceptual
metaphors in European languages and cultures (Trim 2007, 2011).) At the
image schema level, the metaphors tend to be universal and they emerge
unnoticeably in the course of ontogeny and phylogeny.
It seems, then, that context is an unavoidable aspect of metaphorical
meaning making. The conceptual and the contextual aspects of metaphor
8.5 Offline or Online? 177
are inextricably bound together; one cannot exist without the other.
Conceptual metaphor theory cannot simply be just conceptual. It is just as
much contextual. The question in the section title simply does not make too
much sense.
8.5 Offline or Online?

Conceptual metaphor theory started out as, and remained for a long time, an
essentially offline theory of metaphor. What this means is that conceptual
metaphors were regarded as static connections (mappings) between two
domains (such as life is a journey) in long-term memory. Metaphors were
described predominantly as products of the operation of the conceptual system,
rather than the processes that generate the products. In the course of studying
conceptual metaphors, hundreds and perhaps even thousands of such product-
like mental structures were identified and described in detail. With the advent
of the theory of primary metaphors, this situation did not change much in its
character. Despite the fact that primary metaphor theory added new depth and
dimension to our knowledge of conceptual metaphors, it followed and
reinforced the tradition of thinking about metaphor; namely, that conceptual
metaphors, primary metaphors chief among them, are stable and entrenched
products of the conceptual system that arise from human bodily experience. In
a way, the theory of schematicity hierarchies and cascades (see, e.g., David,
Lakoff, and Stickles 2016) does the same: it continues to describe offline
metaphorical structures at the level of image schemas, domains, and frames
in long-term memory. Of course, the new structures are more complex (e.g.,
they can be found on several levels), they integrate aspects of metaphor not
seen as working together before, and they offer elegant explanations of how
certain metaphor-related phenomena work (e.g., how lower level metaphors
inherit information from higher levels).
Critical reaction was bound to happen, and it did. Proponents of conceptual
integration theory suggested that CMT lacks a component that could capture
the dynamic processes that occur in the course of the actual use of metaphors.
To remedy this, they offered the network model (see Chapter 6). Such pro-
cesses include projections that go not just from source to target but also from
target to source, from source and target to blend, and from blend to source and
target. These processes cannot take place between frames and domains in long-
term memory; they must take place in mental spaces. Mental spaces and the
networks of which they form a part are built online at the time of the communi-
cative events that include a metaphorical blend. Similarly, discourse-oriented
work on metaphor showed that in real discourse “metaphoring” is not
happening at the level of frames or domains; these researchers offered the
notions of scenario, scene, and mental space as the level where metaphors are
actually used (see, e.g., Musolff 2006). Still others, like those concerned with
the pragmatic and rhetorical uses of metaphor, complained that CMT has no
way of handling the pragmatic-rhetorical functions of metaphor in natural
discourse (see, e.g., Goatly 1997; Semino 2008; Steen 2008, 2011). It grad-
ually became clear that something important was missing from CMT as a
general theory of metaphor.
The question was: How can we maintain all the achievements of CMT and
at the same time accommodate these criticisms? As I attempted to show in the
previous chapter (Chapter 6), my personal solution to this dilemma is to extend
the schematicity hierarchy to include the level of mental spaces (or scenarios).
Instead of the usual three-way distinction between image schema metaphors,
domain metaphors, and frame metaphors, I suggest a four-way distinction that
includes mental space level metaphors. The new schematicity hierarchy goes
from the most schematic to the least schematic (or most specific) level: to that
of mental spaces. Mental space level metaphors assume all the higher level
ones in a hierarchy. This way, we can turn CMT into a theory that can combine
the idea of an offline schematicity hierarchy and its online functioning in a
unified view of metaphor as both a product and a process.
As a final demonstration of how this might work, consider the example of
capsize again. As mentioned above, a journalist is interviewing Fats Domino
about his life in the wake of Hurricane Katrina. In the course of interpreting
what happened to Domino, the journalist gets to a point where he wants to say
that Domino’s life took an abrupt and bad turn. At the time of the interview all
the visible signs of the consequences of the hurricane are still present, such as
the overturned boats. He writes: “The 2005 hurricane capsized Domino’s
life,. . .” The perceptual property of what he saw then and remembers now
(overturned boats) functions as a contextual factor that activates the image
schema metaphors action is self-propelled motion, purposes are
destinations, and functionality is staying erect. The lack of
functionality (of the boat) amounts to (the boat becoming) an obstacle to
continue motion: that is, to the further image schema (primary) metaphor:
difficulties are obstacles. Since the topic of the conversation is life,
these image schema level metaphors activate the domain level metaphor life
is travel. This matches the primary metaphors and elaborates on them
(Figure 8.2). The life is travel metaphor evokes the more specific meta-
phor living a life is journeying at the frame level. Now the journalist is
in a position to take the journeying frame and to express the contextual
meaning of “abrupt negative change in Domino’s life” through the highly
specific mental space level metaphor: an abrupt negative change in
domino’s life is the capsizing of domino’s boat during his sea
journey.
8.5 Offline or Online? 179
This can be represented diagrammatically as follows:
ACTION IS SELF-PROPELLED MOTION

PURPOSES ARE DESTINATIONS
FUNCTIONALITY/PERSISTING IS STAYING ERECT
DIFFICULTIES ARE OBSTACLES
LIFE IS TRAVEL
LIVING A LIFE IS JOURNEYING
ABRUPT NEGATIVE CHANGE IN DOMINO’S LIFE IS

THE CAPSIZING OF HIS BOAT IN THE COURSE OF
HIS SEA JOURNEY
CONTEXTUAL FACTOR: visible signs of the hurricane – overturned boats
Figure 8.2 Meaning construction for “The 2005 hurricane capsized

Domino’s life,. . .”
The diagram is somewhat misleading, in that it makes it appear as though

the contextual meaning determines or influences which contextual factor is
selected that evokes the image schema metaphors. However, the contextual
meaning does not have to preexist the contextual factor that is selected. It can
actually follow the emergence of a salient contextual factor. In other words, it
seems that in certain circumstances it is the emergence of a salient contextual
factor that can determine or influence what the speaker will want to express
metaphorically. An example for this might be a metaphor used by Matthew
Arnold in his poem Dover Beach (for a detailed analysis, see Kövecses 2015a).
In the poem, Arnold looks on as the sea ebbs and flows, exposing “the naked
shingles of the earth.” Since in the same stanza he writes about “the Sea of
Faith,” we can be reasonably certain that the idea that he is conveying is how
the Christian faith is disappearing in nineteenth-century England, thus leaving
the world vulnerable (“naked”). It may well be that the particular context with
the exposed earth gave him the very idea to express what he did, in which case
we could suggest that it was the physical context (i.e., what he saw) that
evoked the metaphorical idea, rather than the other way around. Probably, we
all remember situations where the context gave us a metaphorical idea, rather
than the idea leading us to some context through which we expressed the
preexisting idea metaphorically.
8.6 Final Words

The extended view assumes much of “standard” CMT, as characterized in
Chapter 1. But the question arises how the extended CMT is related to other
theories of metaphor. In most cases, there are both shared parts and there are
incompatibilities.
As regards conceptual integration theory (an extension of CMT itself ), the
present extended view regards metaphorical blending (i.e., the construction of
metaphor blends) as a process that is happening at the level of mental spaces in
working memory but, at the same time, as a process in the course of which all
the higher level metaphorical structures in long-term memory are utilized.
Metaphorical blending is, in other words, a part of the functioning of a larger
system that is characterized by both compatibilities and incompatibilities
between domains and frames. Blends occur where there are incompatibilities.
Something similar can be said about the relationship between the extended
view and deliberate metaphor theory. Deliberate metaphors that are based on
correlation metaphors assume entire schematicity hierarchies. (However, when
they are based on resemblance, this is clearly not the case.) The former type of
deliberate metaphors are embodied and their comprehension activates the
relevant hierarchy. Unlike the claims by deliberate metaphor theorists, con-
ventionalization of metaphorical meaning does not equal a lack of activation,
as numerous studies indicate.
The extended view shares with structure-mapping theories the idea that
metaphors are sets of systematic mappings between two domains or frames
(Gentner 1983). But while structure-mapping theories primarily study resem-
blance metaphors (metaphorical analogies) that do not involve schematicity
hierarchies, the extended view studies correlation metaphors that do.
The extended view of CMT has even something in common with what
relevance theory has to say about metaphor, in that the inferential processes
employed in relevance theory in metaphorical meaning construction build on
the notion of mutual cognitive environment. This notion comes close to that of
context, as used here. In both views, the production and comprehension of
metaphor is heavily context-bound. Where the two views differ is that rele-
vance theoretical accounts do not accept the existence of conceptual metaphors
(let alone the existence of schematicity hierarchies).
8.6 Final Words 181
But the view of metaphor that is most compatible with the extended view is
that of the dynamic systems view of metaphor, as proposed by Gibbs in
various publications (e.g., Gibbs 2013, 2017c; Gibbs and Cameron 2007).
Dynamic system models make use of a wealth of information as the system of
adaptive behavior moves forward. The system takes in information about the
body (both present and historical), about the conceptualizers who participate in
the communicative situation, the discourse itself, the goals of the participants,
the objects and events present in the situation, and the broader environment.
The dynamic systems view and the extended CMT share the assumption that
the totality of this information and all of our experiences can contribute to
metaphor production and comprehension. To capture this totality, the extended
CMT invokes four types of context: situational, discourse, conceptual-
cognitive, and bodily. It also assumes, probably with the dynamic systems
model, that any of this information and experience can prime the use of
metaphors as discourse progresses. Given the diversity of factors they rely
on, both views can be expected to do well in accounting for metaphorical
creativity and context-sensitivity.
Gibbs (2013: 72–73) regards metaphoric activity “as a dynamic, self-
organized process” and considers metaphor “as an emergent product of multiple
constraints operating along different time-scales. . . .” Let us return to the
example of capsize, as Gibbs (2017c: 331–332) uses this metaphor from the
article about Fats Domino to demonstrate the kinds of forces, or constraints, that
all contribute to the emergence of this metaphor. He provides the following list:
Evolutionary forces to maintain survival and enhance cooperation to
enhance individual and group welfare. Primary metaphors (e.g., persist-
ing is remaining erect) emerge from both evolutionary and experien-
tial forces.
Culturally specific ideas about different topics, such as how best to deal with
difficult life challenges.
Historical forces related to the author’s past experiences as a writer for the
specific forum in which the Fats Domino profile was published.
Social forces that enable speakers to negotiate specific meaning in particular
contexts.
Linguistic factors related to conventions about particular word and gram-
matical uses including, in this case, employing nautical terms to refer to
human experiences.
Immediately salient contextual information such as the presence of the
aftereffects of Hurricane Katrina.
Immediate communicative motivations to state something that is vivid,
memorable, contextually relevant, and succinct regarding a rather complex
abstract idea related to difficulties in one’s life.
Previously stated words that primed the author’s use of certain boating and
water terms that were related to earlier talk about the physical damage to
buildings caused by Hurricane Katrina.
Bodily experiences related to enduring concepts that are relevant for speak-
ing about temporary setbacks in life, such as the idea of being unable to
persist is not being able to remain erect or upright.
Brain and neural activity including that which emerges as brain systems
become coupled during particular interaction between speakers and listeners
or writers and readers.
Let us go over this list of forces and see how the extended CMT would
handle them.
In the extended CMT, evolutionary forces play a role as the shapers of
primary metaphors, and thus are part of the bodily context. The primary
metaphors themselves, like persisting is remaining erect, are part of
the conceptual-cognitive context.
The force that Gibbs calls “culturally specific ideas about different topics”
could probably be accommodated in the system I have proposed as part of the
cultural context – if “culturally specific ideas” mean the knowledge and
experience that members of a speech community share relative to such topics.
The historical forces (in the sense of “the author’s past experiences as a
writer for the specific forum in which the Fats Domino profile was published”)
could probably be accommodated as the personal history of the author within
the conceptual-cognitive context.
The social forces “that enable speakers to negotiate specific meaning in
particular contexts” would seem to fit into what I call the social context within
the more inclusive context type of situational context.
The linguistic factor of conventionally using nautical terms to talk about life
and human experiences in general essentially boils down to how conceptual
metaphors are given linguistic expression in discourse. Here the conceptual
metaphor would be life is a sea journey. This would again be part of the
conceptual-cognitive context, and the expressions would form part of the
mental lexicon that attaches to the conceptual system.
Immediately salient contextual information corresponds to the situational
context (more specifically, the physical environment) in the extended CMT
framework. All the perceivable aftereffects of the hurricane are potential
contextual factors in the use of the metaphor.
Immediate communicative motivation (in the sense of communicative goals)
is not something that is part of any of the four context types, which deal with
actual contentful experiences. (I will say more about this issue below.)
The force of previously stated terms related to the damage caused by the
hurricane comes close to what I called “previous discourse on the same topic” –
8.6 Final Words 183
an aspect of the discourse context. This can include all the boating and water
related words that were used in the press previously.
The force of bodily experiences related to the idea of being unable to
persist is not being able to remain upright could clearly be
accounted for in the extended CMT framework – both as part of the bodily
context (as an experiential correlation) and as part of the conceptual-cognitive
context (as a variant of a primary metaphor).
Finally, the brain and neural activity can be accommodated within extended
CMT as regards the neural correlates of experiential correlations (bodily con-
text), but not as regards the neural couplings between speakers and listeners.
We can now make some further general observations concerning the simi-
larities and differences between the two approaches. As can be seen in light of
the comparison above, certain forces, or constraints, from the dynamic systems
view are not matched or cannot easily be accommodated by the extended CMT
view. This makes the dynamic systems view broader in its scope. It is
interesting to see what causes this imbalance between the two frameworks.
The major goal of the extended view of CMT so far was to capture how
“contentful experience” in various types of context can prime the use of
metaphors in particular discourse situations (see Kövecses 2015a). To this
end, four different “context types” were assumed, including the situational,
discourse, bodily, and conceptual cognitive context. These are all constituted
by various “contextual factors” that correspond to some contentful experience.
The contentful experience can be bodily but also situation-, discourse-, and
(cognitive) representation-based. This last one is what I term “conceptual-
cognitive context” and it simply means that the contentful experience is
already stored (as in the case of conventional conceptual metaphors).
If we look at the forces of the dynamic systems view in relation to capsize,
we can see that all four context types are present on Gibbs’ list. But his list
includes more, such as “the immediate communicative motivation,” that have
no obvious parallel on my list of kinds of contentful experience. The reason,
I believe, is that the missing items can be both behavior-organizing systems
and contentful experiences in the sense in which I am using the term. If we
take “the immediate communicative motivation” as an example, we can see
that communicative motivation is an organizing principle of human (communi-
cative) behavior, rather than a piece of contentful experience, such as the
capsizing of a boat (situational context), what was said about the hurricane
earlier (discourse context), or being unable to get up when seriously sick
(bodily context), or finding the primary metaphor persisting is remaining
erect in the metaphorical conceptual system (conceptual-cognitive context).
Additional organizing principles of (metaphor-related) behavior include the
purposive (telic) component of behavior, the emotion and attitude system, and
so on. For example, it can be observed that when we are angry, we tend to use
different metaphors than when we are happy or sad. The behavior-organizing

systems are often constraints on content(ful experience), but they are not
content itself. They organize and are about content. This is why I did not
include them on my list of contextual factors in my book Where metaphors
come from (Kövecses 2015a). However, I added there:
I did not say much in this book about such important factors as the emotional state of
the speaker/hearer, the attitude of the speaker/hearer to a particular topic, or the purpose
of the use of metaphor, including the purpose of being playful, funny or serious. My
primary concern was with what contentful experiences (summarized as the contextual
factors in this chapter) can prime and thus lead to particular metaphors with a particular
conceptual content (on priming, see later section), and not with the factors that constrain
or delimit their presentation in some way, as is the case with emotions, attitudes, and
purposes. (Kövecses 2015a: 188)
In other words, I make a distinction between constraints, on the one hand, and
factors with a priming effect, on the other. In a way, of course, factors with a
priming effect (i.e., some experiential content) can also be viewed as forces, or
constraints, of a special kind. Gibbs’ forces, or constraints, are largely inclu-
sive of both. In this respect, the dynamic system view provides a more
complete picture of metaphor performance because, in addition to the factors
with a priming effect, it draws attention to the social embeddedness of the use
of metaphors in terms of the social dynamics (between participants of dis-
course and their environment) that is enacted every time we use metaphors
(see, also, Landau 2017 in this connection).
This does not mean, however, that emotions, moods, motivations, purposes,
and so on cannot be used as experiential content. When they are, they can be
priming sources of metaphors (and not behavior-organizing systems). Take the
following situation: Suppose my computer dies unexpectedly several times,
only to work well in between the changes. Suppose, furthermore, I have a very
moody partner whose mood often changes in unpredictable ways. So when my
computer dies on me again next time, I can say: “My computer is as moody as
you are.” Here, mood is used as experiential content that can prime its use as
metaphor. It is the contentful experience of the resemblance between my
computer’s behavior and my partner’s behavior that functions as a prime for
my use of this particular metaphor in the situation.
Overall, it seems to me that the dynamic systems view of metaphor can be
considered, at least in my terms, a predominantly mental-space level theory. It
attempts to capture the phenomenon of metaphor as an online process in
working memory. In this sense, we can suggest that it is a “mental-space-
heavy” theory that does not attribute a great deal of importance to more stable
conceptual structures that might reside in long-term memory, such as frames
and domains. By contrast, the extended CMT I propose sees all these concep-
tual structures as functioning together, integrating what goes on at the mental
8.6 Final Words 185
space level with more or less permanent or stable conceptual structures. This
integration occurs by means of higher-level structures shaping mental spaces,
on the one hand, and mental-space level structures activating higher-level
structures, on the other. Moreover, since the mental spaces are embedded in
information-rich context, the various contextual factors also influence what
goes into the emerging mental spaces. Notice that I am not suggesting than the
extended CMT is “better” than the dynamic systems view. Actually, since my
argument assumes the existence of long-term conceptual structures, if it turns
out that there are no such conceptual structures in the mind, then clearly the
dynamic systems view is to be preferred.
In addition to the question of how the extended CMT view is related to other
theories of metaphor, there are many other pending ones. First, if the extended
CMT and dynamic systems views are compatible, how could they be turned
into a single unified theory? How can their respective advantages be best
utilized and how can their respective disadvantages be eliminated? Second,
how could either one, or the unified version, be turned into a processing model
of metaphor use? Third, how is it possible to devise experiments that work
with the totality of our experiences that the two views attempt to make use of
within their frameworks? Fourth, to continue with the same issue, how can this
bewildering complexity and diversity of information, experience, timescale,
processing, and interaction be turned into a neural theory of metaphor? Can
any computational model take all this into account, including the challenge
that comes from the contextual component? And, finally, how can we study all
this happening in human subjects?
At least to my mind, these are big questions and challenges for the future.
For the time being, however, we should probably aim lower. Let us, first of all,
see if in addition to the linguistic evidence for the schematicity hierarchy we
can find supporting evidence from other modalities. Then, let us see further
specific examples of contextual influence on how linguistic and nonlinguistic
metaphors and the corresponding hierarchies are put to use in actual cases of
metaphorical meaning making. Following that, we can begin to think about the
role and importance of context-induced hierarchies more generally in social
cognition. Given some ideas about that, we can begin to build more accurate
models for what is going on in the physical and social world, in discourse, in
the body, in the conceptual system, and in the brain, when we interact
metaphorically with each other. The “interactional turn” in the study of
metaphor is right before us in its full (?) complexity.
References
Aitchison, Jean. 1987. Words in the Mind. London: Blackwell.

Altarriba, J. and Bauer, L. M. 2004. The distinctiveness of emotion concepts:
A comparison between emotion, abstract, and concrete words. The American
Journal of Psychology, 117, 389–410. http://dx.doi.org/10.2307/4149007
Andor, József. 1985. On the psychological relevance of frames. Quaderni di Semantica,
VI, 2, 212–21.
Bambini V., Gentili, C., Ricciardi, E., Bertinetto, P., and Pietrini, P. 2011. Decomposing
metaphor processing at the cognitive and neural level through functional magnetic
resonance imaging. Brain Research Bulletin. 86(3–4), 203–16.
Barcelona, Antonio (ed.). 2000. Metaphor and Metonymy at the Crossroads. Berlin:
Mouton de Gruyter.
2000. On the plausibility of claiming a metonymic motivation for conceptual
metaphor. In Barcelona, A. (ed.), Metaphor and Metonymy at the Crossroads,
pp. 32–58. Berlin: Mouton de Gruyter.
Barsalou, Lawrence. 1999. Perceptual symbol systems. Behavioral and Brain Sciences,
22, 577–660.
Benczes, Réka, Barcelona, Antonio, and de Mendoza Ibanez, Francisco José Ruiz
(eds.). 2012. Defining Metonymy in Cognitive Linguistics: Towards a Consensus
View. Amsterdam: John Benjamins.
Benedek, András and Veszelszki, Ágnes (eds.). 2017. Virtual Reality – Real Visuality.
Virtual, Visual, Veridical. Frankfurt am Main: Peter Lang.
Borghi, A. M., Binkofski, F., Castelfranchi, C., Cimatti, F., Scorolli, C., and
Tummolini, L. 2017. The challenge of abstract concepts. Psychological Bulletin.
Advance online publication. http://dx.doi.org/10.1037/bul0000089
Boroditsky, Lera. 2001. Does language shape thought? Mandarin and English speakers’
conception of time. Cognitive Psychology, 43, 1–22.
Boroditsky, Lera and Ramscar, Michael. 2002. The roles of body and mind in abstract
thought. Psychological Science, 13(2), 185–89.
Bowdle, B. F. and Gentner, D. 2005. The career of metaphor. Psychological Review,
112(1), 193–216.
Cameron, Lynne. 2003. Metaphor in Educational Discourse. London: Continuum.
2008. Metaphor and talk. In Gibbs, R. (ed.), The Cambridge Handbook of
Metaphor and Thought, pp. 197–211. Cambridge and New York: Cambridge
University Press.
Cameron, Lynne and Maslen, Robert (eds.). 2010. Metaphor Analysis: Research Practice
in Applied Linguistics, Social Sciences and the Humanities. London: Equinox.
186
References 187
2010. The discourse dynamics framework for metaphor. In Cameron, Lynne, and
Maslen, Robert (eds.), Metaphor Analysis: Research Practice in Applied
Linguistics, Social Sciences and the Humanities, pp. 77–96. London: Equinox.
Casasanto, Daniel. 2009. Embodiment of abstract concepts: Good and bad in right and
left handers. Journal of Experimental Psychology: General, 138(3), 351–67.
Charteris-Black, Jonathan. 2004. Corpus Approaches to Critical Metaphor Analysis.
Basingstoke and New York: Palgrave-MacMillan.
2017. Fire Metaphors: Discourses of Awe and Authority. London-New York:
Bloomsbury.
Cienki, Alan. 2007. Frames, idealized cognitive models, and domains. In Geeraerts, D.,
and Cuyckens, H. (eds.), The Oxford Handbook of Cognitive Linguistics,
pp. 170–87. Oxford: Oxford University Press.
Cienki, Alan and Müller, Cornelia. 2008. Metaphor, gesture, and thought. In Gibbs,
Raymond (ed.), The Cambridge Handbook of Metaphor and Thought,
pp. 483–501. New York: Cambridge University Press.
Clark, Herbert. 1996. Using Language. Cambridge: Cambridge University Press.
Clausner, Tim and Croft, William. 1997. Productivity and schematicity in metaphors.
Cognitive Science, 21(3), 247–82.
Coulson, Seana. 2008. Metaphor comprehension and the brain. In Gibbs, R. (ed.),
The Cambridge Handbook of Metaphor and Thought, pp. 177–94. Cambridge:
Cambridge University Press.
Coulson, Seana and Oakley, Todd. 2000. Blending basics. Cognitive Linguistics,
11(3/4), 175–196.
2003. Metonymy and conceptual blending. In Panther, K.-U. and Thornburg, L. (eds.),
Metonymy and Pragmatic Inferencing, pp. 51–79. Amsterdam: John Benjamins.
Croft, William. 1993. The role of domains in the interpretation of metaphors and
metonymies. Cognitive Lingustics, 4, 335–70.
Dancygier, Barbara and Sweetser, Eve. 2014. Figurative Language. New York:
David, Oana, Lakoff, George, and Stickles, Elise. 2016. Cascades in metaphor and
grammar. Constructions and Frames, 8(2), 214–55.
Deignan, Alice. 1995. Collins Cobuild English Guides 7: Metaphor. London:
HarperCollins.
2003. Metaphorical expressions and culture: An indirect link. Metaphor and Symbol,
18(4), 255–71.
2005. Metaphor and Corpus Linguistics. Amsterdam: John Benjamins.
Dirven, René and Pörings, Ralf (eds.). 2002. Metaphor and Metonymy in Comparison
and Contrast. Berlin-New York: Mouton de Gruyter.
Edwards, Derek. 1999. Emotion discourse. Culture and Psychology, 5(3), 271–91.
El Refaie, Elisabeth. 2019. Visual metaphor and embodiment in graphic illness
narratives. New York: Oxford University Press.
Evans, Vyvyan. 2013. Language and Time: A Cognitive Linguistics Approach.
Cambridge: Cambridge University Press.
Fauconnier, Gilles. 1994. Mental Spaces. New York: Cambridge University Press.
2007. Mental spaces. In Geeraerts, D., and Cuyckens, H. (eds.), The Oxford
Handbook of Cognitive Linguistics, pp. 371–76. Oxford: Oxford University Press.
Fauconnier, Gilles and Turner, Mark. 2002. The Way We Think. New York: Basic Books.
188 References
2008. Rethinking metaphor. In Gibbs, Raymond (ed.), The Cambridge Handbook of

Metaphor and Thought, pp. 53–66. New York: Cambridge University Press.
Feldman, Jerome. 2006. From Molecule to Metaphor. Cambridge, MA: MIT Press.
Fillmore, Charles. 1982. Frame semantics. In Linguistic Society of Korea (ed.),
Linguistics in the Morning Calm, pp. 111–35. Seoul: Hanshin.
Forceville, Charles. 1996. Pictorial Metaphor in Advertising. London: Routledge.
2008. Metaphor in pictures and multimodal representations. In Gibbs, Raymond
(ed.), The Cambridge Handbook of Metaphor and Thought, pp. 462–82. New
York: Cambridge University Press.
2016. Pictorial and multimodal metaphor. In N-M. Klug and H. Stöckl, eds., Handbuch
Sprache im multimodalen Kontext. 241–260. Berlin: De Gruyter.
Forceville, Charles and Urios-Aparisi, Eduardo (eds.). 2009. Multimodal Metaphor.
Berlin: Mouton de Gruyter.
FrameNet. https://framenet.icsi.berkeley.edu/fndrupal/frameIndex
Fusaroli, Ricardo and Morgagni, Simone. 2009. Conceptual metaphor theory: 30 years
after. Cognitive Semiotics, 5(1–2).
Gallese, Vittorio and Lakoff, George. 2005. The brain’s concepts. Cognitive
Neuropsychology, 22(3–4), 455–79.
Gelfand, Michele and McCusker, C. 2001. Culture, Metaphor and Negotiation. Handbook
of Cross-Cultural Management. Blackwell Publishers, New York, 292–314.
Gentner, Dedre. 1983. Structure-mapping: A theoretical framework for analogy.
Cognitive Science, 7, 155–70.
Gibbs, Raymond W. 1994. The Poetics of Mind. Cambridge and New York: Cambridge
University Press.
1999. Researching metaphor. In Cameron, Lynne, and Low, Graham (eds.), Researching
and Applying Metaphor, pp. 29–47. Cambridge: Cambridge University Press.
1999. Taking metaphor out of our heads and putting it in into the cultural world.
In Gibbs, R. W. and Steen, G. (eds.), Metaphor in Cognitive Linguistics,
pp. 145–66. Amsterdam: John Benjamins.
2003. Prototypes in dynamic meaning construal. In Gavins, J., and Steen, G. (eds.),
Cognitive Poetics in Practice, pp. 27–40. London: Routledge.
2006. Embodiment and Cognitive Science. Cambridge and New York: Cambridge
University Press.
(ed.). 2008. The Cambridge Handbook of Metaphor and Thought. Cambridge and
New York: Cambridge University Press.
2009. Why do some people dislike conceptual metaphor theory. Journal of Cognitive
Semiotics, 1–2, 14–36.
2013. Metaphoric cognition as social activity: Dissolving the divide between metaphor
in thought and communication. Metaphor and the Social World, 3, 54–76.
(ed.). 2016. Mixing Metaphor. Amsterdam: John Benjamins.
2017a. Metaphor Wars: Conceptual Metaphors in Human Life. New York and
2017b. Seven empirical challenges for cognitive linguistics. Journal of Cognitive
Linguistics: The Journal of the Japanese Cognitive Linguistics Association, 2(3),
25–38.
2017c. The embodied and discourse views of metaphor: Why these are not so different and
how they can be brought closer together. In Hampe, Beate (ed.), Metaphor: Embodied
Cognition and Discourse, pp. 319–35. Cambridge: Cambridge University Press.
References 189
Gibbs, Raymond W. and Cameron, Lynne. 2007. Social-cognitive dynamics of

metaphor performance. Cognitive Systems Research, 9, 64–75.
Gibbs, Raymond W. and Colston, Herbert. 2012. Interpreting Figurative Meaning.
Cambridge and New York: Cambridge University Press.
Goatly, Andrew. 1997. The Language of Metaphors. London: Routledge.
2007. Washing the Brain: Metaphor and Hidden Ideology. Amsterdam: John
Benjamins.
Goldberg, Adele. 1995. Constructions: A Construction Grammar Approach to
Argument Structure. Chicago, IL: University of Chicago Press.
Goossens, Louis. 1990. Metaphtonymy: The interaction of metaphor and metonymy in
expressions for linguistic action. Cognitive Linguistics, 1(3), 323–40.
Grady, Joseph E. 1997a. Foundations of Meaning: Primary Metaphors and Primary
scenes. Ph.D. diss., University of California at Berkeley.
1997b. theories are buildings revisited. Cognitive Linguistics 8, 267–90.
1999. A typology of motivation for conceptual metaphor. In Gibbs, R., and Steen, G.
(eds.), Metaphor in Cognitive Linguistics, pp. 79–100. Amsterdam: John Benjamins.
Grady, Joseph. 2005. Primary metaphors as inputs to conceptual integration. Journal of
Pragmatics, 37, 1595–614.
Grady, Joseph and Johnson, Christopher. 2002. Converging evidence for the notions of
subscene and primary scene. In Dirven, René, and Pörings, Ralf (eds.), Metaphor
and Metonymy in Comparison and Contrast, pp. 533–54. Berlin-New York:
Mouton de Gruyter.
Grady, Joseph, Oakley, Todd, and Coulson, Seana. 1999. Blending and metaphor. In
Gibb, Raymond, and Steen, Gerard (eds.), Metaphor in Cognitive Linguistics,
pp. 101–24. Amsterdam: Benjamins.
Hampe, Beate. 2005. Image schemas in cognitive linguistics: Introduction.
In B. Hampe, with J. Grady (eds.), From Perception to Meaning: Image Schemas
in Cognitive Linguistics, pp. 1–12. Berlin: Mouton de Gruyter.
Holland, Dorothy and Quinn, Naomi (eds.). 1987. Cultural Models in Language and
Thought. New York: Cambridge University Press.
Johnson, Mark. 1987. The Body in the Mind. Chicago: The University of Chicago Press.
Katz, A. N., Cacciari, C., Gibbs, R. W. and Turner, M. 1998. Figurative Language and
Thought. New York and Oxford: Oxford University Press.
Kimmel, Michael. 2010. Why we mix metaphors (and mix them well): Discourse
coherence, conceptual metaphor, and beyond. Journal of Pragmatics, 42,
97–115.
Kolodny, Annette. 1975. The Lay of the Land: Metaphor as Experience and History in
American Life and Letters. Chapel Hill: The University of North Carolina Press.
1984. The Land Before Her: Fantasy and Experience of the American Frontiers,
1630–1860. Chapel Hill: The University of North Carolina Press.
Kövecses, Zoltan. 1986. Metaphors of Anger, Pride, and Love: A Lexical Approach to
the Study of Concepts. Amsterdam: John Benjamins.
1990. Emotion Concepts. Berlin and New York: Springer-Verlag.
1995a. American friendship and the scope of metaphor. Cognitive Linguistics 6–4,
315–46.
1995b. Anger: Its language, conceptualization, and physiology. In Taylor, J., and
MacLaury, R. (eds.), Language and the Cognitive Construal of the World,
pp. 181–96. Berlin: Gruyter.
190 References
2000a. The scope of metaphor. In Barcelona, Antonio (ed.), Metaphor and

Metonymy at the Crossroads, pp. 79–92. Berlin: Walter de Gruyter.
2000b. Metaphor and Emotion. Cambridge and New York: Cambridge University Press.
2000c. American English: An Introduction. Peterborough, Canada: Broadview Press.
2002/2010. Metaphor: A Practical Introduction. (1st edition 2002, 2nd edition
2010). New York: Oxford University Press.
2005. Metaphor in Culture: Universality and Variation. Cambridge and New York:
2006. Language, Mind, and Culture: A Practical Introduction. Oxford and New
York: Oxford University Press.
2008. Metaphor and emotion. In Gibbs, R. (ed.), The Cambridge Handbook of
Metaphor and Thought, pp. 380–96. New York: Cambridge University Press.
2010a. Metaphor: A Practical Introduction. 2nd edition. New York: Oxford
University Press.
2010b. A new look at metaphorical creativity in cognitive linguistics. Cognitive
Linguistics, 21(4), 663–97.
2011a. Methodological issues in conceptual metaphor theory. In Handl, S., and
Schmid, H.-J. (eds.), Windows to the Mind: Metaphor, Metonymy and Conceptual
blending, pp. 23–39. Berlin/New York: Mouton de Gruyter.
2011b. Recent developments in metaphor theory: Are the new views rival ones?
Review of Cognitive Linguistics, 9(1), 11–25.
2011c. The biblical story retold: A cognitive linguistic perspective. In Brdar, M.,
Gries, S., and Fuchs, M. (eds.), Cognitive Linguistics: Convergence and
Expansion, pp. 325–54. Amsterdam: Benjamins.
2013. The metaphor-metonymy relationship: Correlation metaphors are based on
metonymy. Metaphor and Symbol, 28(2), 75–88.
2015a. Where Metaphors Come From, Reconsidering Context in Metaphor. Oxford
and New York: Oxford University Press.
2015b. Surprise as a conceptual category. Review of Cognitive Linguistics, 13–2, 270–90.
2015c. Metaphor and emergentism. In MacWhinney , Brian, and O’Grady, William
(eds.), The Handbook of Language Emergence, 147–162. John Wiley and Sons.
2017a. A radical view of the literal-figurative distinction. In Benedek, A., and
Veszelszki, Á. (eds.), Virtual Reality – Real Visuality: Virtual, Visual, Veridical,
pp. 17–28. Frankfurt am Main: Peter Lang.
2017b. Levels of metaphor. Cognitive Linguistics, 28–2, 321–47.
2018. The power (and problem) of money. Society and Economy 40(3), 365–76.
2019. Idioms of money – In a new light. In Duda, B., Kieltyka, R., and Konieczna, E.
(eds.), Culture, Cognition, Discourse and Grammar, pp. 21–32. Berlin: Peter Lang.
In press. Sensing the city: Budapest through its metaphors. In R., Digonnet and S.,
Beligon eds. Manifestations sensorielles des urbanités contemporaines, Berlin:
Peter Lang.
Kövecses, Zoltán and Radden, Günter. 1998. Metonymy: Developing a cognitive
linguistic view. Cognitive Linguistics, 9(7), 37–77.
Kövecses, Zoltán, Ambrus, Laura, Hegedűs, Dániel, Imai, Ren, and Sobczak, Anna.
2019. The lexical vs. corpus-based method in the study of metaphors.
In Bolognesi, M., Brdar, M., and Despot, K. (eds.), Metaphor and Metonymy in
the Digital Age: Theory and methods for building repositories of figurative
language, pp. 149–173. Amsterdam: John Benjamins.
References 191
Lakoff, George. 1987. Women, Fire, and Dangerous Things: What Categories Reveal
About the Mind. Chicago: The University of Chicago Press.
1990. The invariance hypothesis: Is abstract reason based on image schemas?
Cognitive Linguistics, 1, 39–74.
1993. The contemporary theory of metaphor. In Ortony, A. (ed.), Metaphor and
Thought. Second edition, pp. 202–51. Cambridge and New York: Cambridge
University Press.
1995. Metaphor, morality, and politics, or, why conservatives have left liberals in the
dust. www.wwcd.org/issues/Lakoff.html (First published in Social Research 62
(2).)
1996. Moral Politics: How Liberals and Conservatives Think. Chicago: The
University of Chicago Press.
2008. The neural theory of metaphor. In Gibbs, Raymond (ed.), The Cambridge
Handbook of Metaphor, pp. 17–38. Cambridge and New York: Cambridge
University Press.
Lakoff, George and Johnson, Mark. 1980. Metaphors We Live By. Chicago:
The University of Chicago Press.
1999. Philosophy in the Flesh. New York: Basic Books.
Lakoff, George and Kövecses, Zoltán. 1987. The cognitive model of anger inherent
in American English. In Holland, D., and Quinn, N. (eds.), Cultural Models in
Language and Thought, pp. 195–221. New York and Cambridge: Cambridge
University Press.
Lakoff, George and Turner, Mark. 1989. More Than Cool Reason. Chicago:
The University of Chicago Press.
Landau, Mark. 2017. Conceptual Metaphor in Social Psychology: The Poetics of
Everyday Life. New York and London: Routledge.
Langacker, Ronald. 1987. Foundations of Cognitive Grammar. Stanford: Stanford
University Press.
2008. Cognitive Grammar: A Basic Introduction. New York: Oxford University
Press.
Low, G., Todd, Z., Deignan, A. and Cameron, L. (eds). 2010. Researching and
Applying Metaphor in the Real World. Amsterdam: Benjamins.
Müller, Cornelia. 2008. Metaphors Dead and Alive, Sleeping and Waking: A Dynamic
View. Chicago: The University of Chicago Press.
Musolff, Andreas. 2001. Political imagery of Europe: A house without exit doors?
Journal of Multilingual and Multicultural Development, 21(3), 216–29.
2004. Metaphor and Political Discourse. Analogical Reasoning in Discourse About
Europe. London: Palgrave Macmillan.
2006. Metaphor scenarios in public discourse: Metaphor and Symbol, 21(1),
23–38.
2016. Political Metaphor Analysis: Discourse and Scenarios. Bloomsbury.
Narayanan, Srini. 1999. Moving right along: A computational model of metaphoric
reasoning about events. Proceedings of the National Conference on Artificial
Intelligence (AAAI '99), Orlando, Florida, July 18–22, 1999, pp. 121–28,
AAAI Press.
Pragglejaz Group. 2007. MIP: A method for identifying metaphorically used words in
discourse. Metaphor and Symbol, 22(1), 1–39.
192 References
Quinn, Naomi. 1991. The cultural basis of metaphor. In Fernandez, J. (ed.), Beyond
Metaphor: The Theory of Tropes in Anthropology. pp. 56–93. Stanford: Stanford
University Press.
Radden, Günter. 2002. How metonymic are metaphors? In Dirven, René and Pörings,
Ralf (eds.), Metaphor and Metonymy in Comparison and Contrast, pp. 407–33.
Berlin-New York: Mouton de Gruyter.
Radden, Günter and Kövecses, Zoltán, 1999. Towards a theory of metonymy. In
Panther, Uwe-Klaus, and Radden, Günter (eds.), Metonymy in Language and
Thought, 17–59. Amsterdam: John Benjamins.
Rakova, Marina. 2003. The Extent of the Literal: Metaphor, Polysemy and Theories of
Concepts. Palgrave Macmillan.
Reddy, Michael. 1979. The conduit metaphor – A case frame conflict in our language
about language. In Ortony, A. (ed.), Metaphor and Thought, pp. 284–324.
Rosch, Eleanor. 1978. Principles of categorization. In Rosch, E., and Lloyd, B. B. (eds.),
Cognition and Categorization, pp. 27–48. Hillsdale, NJ: Lawrence Erlbaum.
Ruiz de Mendoza, Francisco. 1998. On the nature of blending as a cognitive
phenomenon. Journal of Pragmatics, 30, 259–74.
Ruiz de Mendoza, Francisco and Galera, Alicia. 2014. Cognitive Modeling.
A Linguistic Perspective. Amsterdam: John Benjamins.
Ruiz de Mendoza, F. J. and Mairal, R. 2007. High-level metaphor and metonymy in
meaning construction. In Radden, G., Köpcke, K.-M., Berg, T., and Siemund P.
(eds.), Aspects of Meaning Construction, pp. 33–51. Amsterdam/Philadelphia:
John Benjamins.
Searle, John. 1993. Metaphor. In Ortony, A. (ed.), Metaphor and Thought, pp. 92–123.
Cambridge and New York: Cambridge University Press.
Semino, Elena. 2008. Metaphor and Discourse. Cambridge: Cambridge University Press.
Semino, Elena and Demjén, Zsófia (eds.). 2017. The Routledge Handbook of Metaphor
and Language. London: Routledge.
Sharifian, Farzad. 2011. Cultural Conceptualizations and Language. Theoretical
Framework and Applications. Amsterdam: Benjamins.
(ed.). 2015. The Routledge Handbook of Language and Culture. Milton Park: Routledge.
(ed.). 2017a. Advances in Cultural Linguistics. Singapore: Springer.
2017b. Cultural Linguistics. Amsterdam: John Benjamins.
Shen, Yeshayahu and Balaban, Noga. 1999. Metaphorical (in)coherence in discourse.
Discourse Processes, 28(2), 139–53.
Sinha, Chris. 2007. Cognitive linguistics, psychology, and cognitive science. In
Geeraerts, D., and Cuyckens, H. (eds.), The Oxford Handbook of Cognitive
Linguistics, pp. 1266–94. New York: Oxford University Press.
Sperber, Dan and Wilson, Deidre. 1986/1995. Relevance: Communication and
Cognition. Cambridge, MA: Blackwell.
Steen, Gerard. 2008. The paradox of metaphor: Why we need a three-dimensional
model of metaphor. Metaphor and Symbol, 23(4), 213–41.
2011. The Contemporary Theory of Metaphor – Now New and Improved! Review
of Cognitive Linguistics, 9(1):26–64.
2013. Deliberate metaphor affords conscious metaphorical thought. www.academia
.edu/363751/Deliberate_metaphor_affords_conscious_metaphorical_thought
References 193
Stefanowitsch, Anatol. 2006. Words and their metaphors. In Stefanowitsch, Anatol, and
Gries, Stefan Th. (eds.), Corpus-Based Approaches to Metaphor and Metonymy,
pp. 64–105. Berlin: Mouton de Gruyter.
Sullivan, Karen. 2006. Frame-based constraints on lexical choice in metaphor. 32nd
Annual Meeting of the Berkeley Linguistics Society (BLS), Berkeley, CA, 2006.
https://ssrn.com/abstract=1492026
2013. Frames and Constructions in Metaphoric Language. Amsterdam: John
Benjamins.
2016. Integrating constructional semantics and conceptual metaphor. Constructions
and Frames, 8(2), 141–65.
Sweetser, Eve. 1990. From Etymology to Pragmatics: Metaphorical and Cultural
Aspects of Semantic Structure. Cambridge: Cambridge University Press.
Taylor, John. 1989. Linguistic Categorization: Prototypes in Linguistic Theory.
Oxford: Clarendon Press.
Taylor, John R. and MacLaury, Robert (eds.). 1995. Language and the Cognitive
Construal of the World. Berlin: Mouton de Gruyter.
Taylor, John and Mbense, Thandi. 1998. Red dogs and rotten mealies: How Zulus talk
about anger. In Athanasiadou, A., and Tabakowska, E. (eds.), Speaking of
Emotions, pp. 191–226. Berlin: Mouton de Gruyter.
Traugott, E. C. and Dasher, R. B. 2002. Regularity in Semantic Change. Cambridge:
University of Cambridge Press.
Trim, Richard. 2007. Metaphor Networks: The Comparative Evolution of Figurative
Language. Houndmills, Basingstoke: Palgrave Macmillan.
2011. Metaphor and the Historical Evolution of Conceptual Mapping. Houndmills,
Basingstoke: Palgrave Macmillan.
Turner, Mark. 1996. The Literary Mind. New York: Oxford University Press.
Turner, Mark and Fauconnier, Gilles. 2000. Metaphor, metonymy, and binding.
In Barcelona, Antonio (ed.), Metaphor and Metonymy at the Crossroads,
pp. 133–45. Berlin: Walter de Gruyter.
Van Dijk, Teun. 2009. Society and Discourse: How Social Contexts Influence Text and
Talk. Cambridge: Cambridge University Press.
Yu, Ning. 1995. Metaphorical expressions of anger and happiness in English and
Chinese. Metaphor and Symbolic Activity, 10, 59–92.
1998. The Contemporary Theory of Metaphor in Chinese: A Perspective from
Chinese. Amsterdam: John Benjamins.
2002. Body and emotion: Body parts in Chinese expression of emotion. Pragmatics
& Cognition, 10(1), 341–67.
Zinken, Jörg. 2007. Discourse metaphors: The link between figurative language and
habitual analogies. Cognitive Linguistics, 18(3), 445–66.
Index
Aitchison, 143 and the “dynamic systems view,” 181–183

Altarriba, 28, 170 and levels of schematicity, 104–105
Andor, 46 local and global, 100
metaphors and blends in context, 143–145
Balaban, 122, 132 and metaphorical creativity, 105–107
Bambini, 87, 147 and “mutual cognitive environment,” 94–95
Barcelona, 34–35, 38, 44 priming and context, 102–103
Barsalou, 35–36, 56 types of, 95–100
Benczes, 35 and visual metaphor, 107–115
blending, 78, 86, 120, 133–149, 158, 180 correlation metaphor, 34–49, 91, 136, 160,
in the multilevel view, 145 162–164, 171–172, 180
Borghi, 28, 170 Coulson, 87, 143–144
Boroditsky, 24, 70, 87, 100, 102, 104 creativity, metaphorical, 11, 20, 92, 105–106,
Bowdle, 159 115, 121, 147
Croft, 36, 56, 173
Cameron, 10–11, 17, 35, 87, 117, 154, 181
Casasanto, 70, 87, 99, 102, 104 Dancygier, 48, 52, 70, 174–175
Charteris-Black, 3–4 Dasher, 41
Cienki, 8, 53, 107 David, 70, 177
Clark, 95, 100 Deignan, 10, 18, 44, 61, 86, 89, 99, 117
Clausner, 56, 173 Dirven, 35, 44
Colston, 17, 70, 89, 93, 100, 102, 104 discourse, 117, 166–168, 173, 177, 181–185
comprehension (of metaphor), 20, 23–27, 93, as context, 119–120
98–99, 119, 153 and mental space, 118
and deliberate metaphor, 180 and metaphor, 122–133
of mixed metaphors, 130–133 metaphor, 154–159
of visual metaphors, 115 and mixed metaphor, 122–133
concept and pragmatic-narrative functions, 120–122
abstract, 7, 12, 56, 74, 136–137, 169–173 domain, 50–70
concrete versus abstract, 22–29, 32–33 versus frame, 53–54
construal, 32–33, 73, 100–102, 170
figurative, 33, 170 El Refaie, 8, 107–108
content Edwards, 121
emotion and experiential, 183–184 elaboration, 20, 57, 59, 73, 105–107, 118, 121
experiential, 68, 103–104, 116, 151, 184 as specialization, 35–39, 49
ontological, 32, 169–171 embodiment, 20, 33–35, 56–57, 68, 87, 93,
context (and metaphor), 11, 20, 57, 68–70, 100, 116, 162, 170–172, 180
89–90, 93–117, 119, 167, 175–177 emergence See metaphor emergence
choice of metaphor, 150–151 emotion, 12, 15–16, 28, 36–39, 77, 119
and conceptual pathways, 103–104 and blends, 137–138
and contextual factors, 164–166, 178–180 and embodiment, 171
and “current discourse space,” 94 and experiential content, 183–184
deliberate, 73–77, 91, 154, 158, 164, 180 and schematization, 172
194
Index 195
entailment See inference Lakoff and Johnson, 1, 9, 17, 33–37, 51,

Evans, 158 59–60, 64, 74, 86, 100, 108, 134, 136,
experience 162–163, 173, 176
contentful, 95, 182–184 Langacker, 9, 24, 29, 32, 35–36, 51, 53–54, 56,
differential, 95–96 82, 90, 94, 102
levels
Fauconnier, 50, 54, 56, 65, 73, 120, 133–134, of metaphor, 87, 89
138, 143–144, 148 ontological, 164–165, 168, 170–171
Feldman, 87 Low, 18
Fillmore, 35–36, 53, 87
Forceville, 8, 107–108 Mairal, 37
frame, 50–70 mappings, 2–4, 12
versus domain, 53–54 from concrete to abstract, 5–6
Fusaroli, 17 direction of, 5, 23, 32, 147
and primary metaphors, 62
Galera, 69 in schematicity hierarchies, 67, 90
Gallese, 87, 147, 162 Maslen, 17
Gelfand, 99 McCusker, 99
generalization (schematization), 35–39, 43, 46, meaning, 116–117, 164–165
48–49 contextualized vs. decontextualized, 158
Gentner, 159, 180 delimiting literal meaning, 23–26
Gibbs, 1–2, 11, 17, 19, 22, 35, 70, 86, 89, 93, kinds of meaning in metaphor, 158
100, 102, 121, 123, 130, 154, 162, 176, literal, 22–29, 32–33
181–184 meaningfulness, metaphorical, 53, 56, 68,
Goatly, 11, 98–99, 117, 178 79, 117, 157, 164–165
Goldberg, 9 metaphorical meaning in discourse, 166–168
Goosens, 36 in mixed metaphors, 130–131
Grady, 5, 9, 34, 43, 45–47, 50–51, 56, 59, 62, of target domain element, 124
100, 108, 136, 140, 146, 157–159, 171 meaning construction See meaning making
grammar (and metaphor), 10 meaning focus, 5, 14, 61, 64, 91, 126, 132,
grounding (of conceptual metaphors), 8–9, 116 139–140
meaning making (or meaning construction),
Hampe, 53 94–95, 100–102, 142, 144, 148, 158, 167,
Holland, 173 176, 180, 185
memory, 54, 63, 68–70, 90–92, 98–99, 108,
idioms, metaphorical, 44, 78–80, 88, 91 117, 122, 145, 148–149, 164–165, 175,
image schema (and metaphor), 9, 50–58, 177, 184
65–70, 74–75, 84, 90, 105, 117, 145, 147, metaphor
149, 153–154, 157, 160–164, 167–168, compound, 9
174–178 conceptual, 1–3
inference (or entailment), metaphorical, 4, 7, 9, conceptual metaphor theory, 17–21
69, 117, 134 and the construction of reality, 6–7
integration, conceptual, 4, 12, 65, 78, 102, 118, critique of conceptual metaphor theory, 17–18
120, 133–149, 158, 173, 177, 180 cultural, 162–164
deep, 71, 73, 77, 91
Johnson, Ch., 35, 43, 45, 47 definition of conceptual metaphor, 1–2
Johnson, M., 9, 35, 53–54, 56, 74, 173 “embodied” versus “discourse,” 154–156
kinds of, 71, 91, 154, 158–162
Katz, 22 and metonymy, 34–49
Kimmel, 123–126 mixed, 122–133
Kolodny, 96 neural theory of, 10
pervasiveness of, 2
Lakoff, 5, 8–10, 13, 15–16, 23, 28, 35–38, primary, 9–11, 14, 19–20, 34–35, 43, 45–48,
50–51, 53, 56, 69, 71–73, 85, 87, 99, 105, 62, 91, 100, 109, 111–113, 137, 155, 157,
134, 147, 158, 162, 173–174, 177 164, 173, 176–178, 181–183
196 Index
metaphor (cont.) schematization, 35–40, 42–44, 49, 102, 137,

and relevance theory, 180–181 172–173
and structure-mapping, 180 of prospective source, 40
and thought, 6 of prospective target, 42
superficial, 71, 73, 91 Searle, 23
visual, 71, 80–86, 92, 107–115 Semino, 11, 87, 98, 103, 117, 178
metaphor emergence, 8, 11, 19, 25, 34–49, 77, Sharifian, 163
99, 171–173 Shen, 122, 132
metaphor networks, 11–17 Sinha, 95
method (in the study of metaphor), 86–90 smell as target domain, 29–32
corpus-based, 3, 10, 44, 86–90 space, mental, 50–70, 76–77, 83, 86, 89–92,
lexical, 3, 18, 60, 69, 87–90 104, 117, 147–148, 164, 174, 185
metonymy, 22, 25, 32, 77, 102, 111, 125, 134, and conceptual integration, 133–134, 145
139, 143, 147, 172–173 definition of, 54
and metonymization, 172–173 and meaning, 148, 152
Morgagni, 17 and mixed metaphor, 122, 133
Müller, 8, 23, 107 and nonuniversality, 163
multilevel view of CMT, 67–70 uses of metaphor at the mental space level,
consequences of, 70–92, 164–166 118–122
multimodality (of metaphors), 7 specialization (or elaboration), 35–39
Musolff, 11, 50, 57–58, 64, 69, 87, 122, Sperber, 94
162–163, 173, 178 Steen, 18, 71, 73, 77, 117, 154, 164, 178
Stefanowitsch, 86
Narayanan, 87 Stickles, 70, 177
structures, conceptual (involved in metaphor),
Oakley, 143–144 45, 50–58, 68, 70, 90–91, 145, 147, 149,
158, 164–165, 173–175, 184
pathway, conceptual, 101, 103–104, 116, styles, differential cognitive, 95–96
151–154, 159–161, 166, 168 Sullivan, 9–10, 41, 54, 87
Pörings, 35, 44 Sweetser, 25, 39, 48, 52, 54, 70, 74,
priming, 24, 102–105, 115–117, 119, 156, 166, 174–175
184 systems, dynamic (view of metaphor), 155
processing (of metaphor), 167–168, 185 compared with extended CMT, 181–185
of mixed metaphors, 130–132
production (of metaphor), 20, 69, 93–94, Taylor, 11, 35, 44
97–104, 167, 180 Traugott, 41
Trim, 12, 176
Quinn, 154, 173 Turner, 16, 50, 54, 65, 73, 105, 120, 133–134,
138, 142–144, 148, 158, 173
Radden, 35–36, 38, 44–46, 48, 140
Rakova, 23 unidirectionality (of metaphor), 5, 32–33
Ramscar, 100, 102 Urios-Aparisi, 107
Reddy, 15
resemblance metaphor, 34–35, 48, 115, 153, Van Dijk, 144
158, 160, 162–164, 180 variation (of metaphor), 11, 50, 89, 95, 107
Rosch, 35–36, 51
Ruiz de Mendoza, 35, 37, 69 Wilson, 94
schematicity hierarchy, 50–92, 103–107, 119, Yu, 11, 97

130, 151–152, 154, 156, 158–161, 163,
168, 178, 185 Zinken, 154, 158

Extended Conceptual Metaphor Theory

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Extended Conceptual Metaphor Theory

Uploaded by

Copyright:

Available Formats

Extended Conceptual Metaphor Theory

The phenomenon known as metaphor is an extremely complex mental event –

zoltán kövecses is Emeritus Professor at Eötvös Loránd University.

Cambridge University Press is part of the University of Cambridge.

List of Figures page viii

1 A Brief Outline of “Standard” Conceptual Metaphor

3.1 Schematization of the prospective source domain page 40

In a way, this is a very personal book. It is personal in the sense that in

Structure of the Book

1.1 The Pervasiveness of Metaphors

1.2 A More Technical Deﬁnition of Conceptual Metaphors

metaphors, ﬁrst, as derived by the lexical method, then, as derived by corpus

We can ﬁnd several of these linguistic metaphors in corpus-oriented work as

Given such examples, the following set of correspondences, or mappings, can

1.3 Mapping Further Knowledge

1.4 Mappings Go from Concrete to Abstract Domains

journeys metaphorically as life, ﬁre as anger, or buildings as theories. We

1.5 Metaphors in Thought

1.6 Metaphor and the Construction of Reality

1.7 Multimodality of Metaphors

gestures, visual representations (such as cartoons), visual arts (such as

1.8 The Grounding of Conceptual Metaphors

conceptual metaphor intensity is heat can then be regarded as grounded

1.9 Primary and Compound Metaphors

1.10 Image Schemas and Metaphors

1.11 Metaphor and Grammar

devised an argument,” “The lawyer devised a house,” “The carpenter devised

1.12 The Neural Theory of Metaphor

1.13 Universality and Variation

1.14 Context and Metaphor

1.15 Metaphor Networks

Different Aspects of a Single Generic Concept

As we can see, the highest-level metaphor here is related to the overarching

A Single Aspect of Several Different Speciﬁc-Level Concepts

Several Different Aspects of a Single Speciﬁc-Level Concept

conceptually partakes of a number of different metaphor systems. Since

The interactions are conceptualized as “economic” exchanges because people

The Target Concepts Form a Hierarchical System of Concepts

categories are often graded internally – a conceptualization that can lead to

1.16 Critique of CMT

1.17 The Present State of CMT

 why many conceptual metaphors are near-universal or potentially universal;

1.18 Extending CMT

Most metaphor researchers accept, either explicitly or implicitly, a distinction

conceptualized as basic concrete experiences. But if this is the case and it is

2.1 Delimiting the Literal

conceptual metaphor” to establish what is literal. The criterion seems to be the

have a series of metonymies at work here. As another example, consider the

2.2 So What Exactly Is Literal Meaning?

abstractions but also in comprehending the concrete, physical aspects of the

The Radical Claim

material in a ﬁgurative way to comprehend and construct abstractions without

2.3 Smell as a Target Domain

The three cases conceptualize the experience of smell as the existence of a

 ADJ. overpowering, pervasive, pungent, rich, sharp, strong There was

The cognitive-construal part of abstract concepts is essentially ﬁgurative

Primary metaphors form the foundation of conceptual metaphor theory.1 They

3.1 A Brief Note on the Conceptual System

3.2 Metaphor or Metonymy?

Case 1: A Frame Element Becomes the Source

happiness, and an increase in body heat is an important ingredient of our

then generalized, or schematized, to the concept of closeness. The words

ELEMENT 1 metonymy ELEMENT 2

Figure 3.1 Schematization of the prospective source domain

why many conceptual metaphors are near-universal or potentially universal;

ADJ. overpowering, pervasive, pungent, rich, sharp, strong There was